understanding the rural - urban reading gap · understanding the rural-urban reading gap catalogue...

Education, skills and learningResearch papers

Understanding the rural - urbanreading gap

Catalogue no. 81-595-MIE2002001

Statistics StatistiqueCanada Canada

Human Resources Développement des ressourcesDevelopment Canada humaines Canada

Council of Ministers of Education, CanadaConseil des ministres de l’Éducation (Canada)



Fernando Cartwright and Mary K. AllenStatistics Canada

Published by authority of the Minister responsible for Statistics Canada

© Minister of Industry, 2002

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system ortransmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwisewithout prior written permission from Licence Services, Marketing Division, Statistics Canada, Ottawa,Ontario, Canada K1A 0T6.

November 2002


Frequency: Irregular

ISSN 1704-8885ISBN 0-662-33191-5

Ottawa

La version française de cette publication est disponible sur demande (no 81-595-MIF2002001 au catalogue).

Statistics Canada

Human Resources Development Canada

Council of Ministers of Education, Canada

This paper represents the views of the authors and does not necessarily reflect the opinions of Statistics Canada,Human Resources Development Canada or the Council of Ministers of Education, Canada.

How to obtain more information

Specific inquiries about this product and related statistics or services should be directedto: Client Services, Culture, Tourism and the Centre for Education Statistics, StatisticsCanada, Ottawa, Ontario, K1A 0T6 (telephone: (613) 951-7608; toll free at 1 800 307-3382;by fax at (613) 951-9040; or e-mail: [email protected]).

For information on the wide range of data available from Statistics Canada, you cancontact us by calling one of our toll-free numbers. You can also contact us by e-mail or byvisiting our Web site.

National inquiries line 1 800 263-1136National telecommunications device for the hearing impaired 1 800 363-7629E-mail inquiries [email protected] site www.statcan.ca

Ordering information

This product, Catalogue No. 81-595-MIE2002001, is available on the Internet for free.Users can obtain single issues at: http://www.statcan.ca/cgi-bin/downpub/studiesfree.cgi.

Standards of service to the public

Statistics Canada is committed to serving its clients in a prompt, reliable and courteousmanner and in the official language of their choice. To this end, the Agency has developedstandards of service which its employees observe in serving its clients. To obtain a copyof these service standards, please contact Statistics Canada toll free at 1 800 263-1136.

3

Understanding the rural-urban reading gap


Acknowledgements

We would like to thank the students, parents, teachers and principals who gave oftheir time to participate in the 2000 OECD PISA study and the Youth in TransitionSurvey. The support for this Federal-Provincial collaborative project provided bymembers of the PISA-YITS Steering Committee and by the coordinators in eachparticipating Ministry or Department of Education during all steps of the study isgratefully acknowledged.

The report has benefited from the input and comments of reviewers, especiallythose in provincial Ministries and Departments of Education, the Council of Ministersof Education Canada, Human Resources Development Canada and StatisticsCanada. The staff of the Centre for Education Statistics at Statistics Canada, whosetireless efforts ensured a high standard of quality, are appreciated. A very specialthank you is extended to Danielle Baum for her indispensable help in preparing themanuscript for publication.

Note of appreciation

Canada owes the success of its statistical system to a long-standing co-operativeeffort involving Statistics Canada, the citizens of Canada, its businesses, governmentsand other institutions. Accurate and timely statistical information could not beproduced without their continued co-operation and good will.

4



Table of Contents

Summary 5

1. Introduction 6

2. Urban students outperform rural students in reading 9

3. How do rural and urban students and schools differ? 12

4. What rural characteristics lie behind the difference in

reading performance? 18

5. Schools matter 21

Appendix A: Tables 24

Appendix B: Definitions and concepts 49

Appendix C: Analytical methodology 55

Appendix D: Survey concepts, methodology and data quality 66

5Catalogue no. 81-595-MIE2002001

SummaryStudents from urban schools in Canada performed significantly better in readingthan students from rural schools, according to the Programme for International StudentAssessment. The rural-urban reading gap was particularly large in Newfoundlandand Labrador, Prince Edward Island, New Brunswick and Alberta.

Students in rural schools in Alberta, while not performing as well as theirurban counterparts, had reading scores above the national average and better thanurban students in some other provinces.

This study uses data from the Youth in Transition Survey (YITS) and theProgramme for International Student Assessment (PISA) to examine the differencein reading performance between students in rural and urban schools and to investigatewhy the rural-urban reading gap exists in some provinces.

The study found that rural students were more likely than urban students tocome from families with lower socio-economic backgrounds. The parents of ruralstudents tend to be less well-educated and less likely to be employed in professionaloccupations, such as doctors, lawyers and bankers. These differences, however, donot explain the gap in performance between rural and urban students. Even if onewere to compare rural and urban students whose parents had the same level ofeducation and the same occupation, the reading difference would still remain.

Moreover, the rural-urban gap cannot be attributed to differences in rural andurban schools because, for the most part, rural and urban schools are much thesame. In fact, Canadian students ranked high internationally, in part, because thereare few significant differences between Canadian schools overall.1

Instead, this study shows that the difference between rural and urban readingperformance is most strongly related to community differences. Relative to theurban communities, rural communities were characterised by lower levels ofeducation, fewer jobs, and jobs that were, on average, lower earning and less likelyto require a university degree. The rural-urban reading differences are linked tocommunity differences in levels of adult education and the nature of work. Thecommunity characteristics are based on both the education and job level of theparents of all of the school’s 15-year-olds, and on the educational and occupationalcharacteristics of the adult population of the school’s municipality.

6



While family background is important, it is the community background inwhich students learn that explains the rural-urban reading gap. That is, the child ofsomeone in a professional occupation will likely perform well in either an urban orrural school, but will likely perform even better in an urban community. In fact,every child will likely do better in an urban community because of the nature of theurban labour market and the overall higher levels of education among the adults. Itis important to note that, because changes in school factors would affect both ruraland urban students, they are not likely to reduce the rural-urban gap.

Because changes that might reduce the reading gap between rural and urbancommunities are not short-term measures, the study examined a variety of schoolcharacteristics that were not reported at the highest levels in rural schools in order toidentify which of these characteristics has a strong relationship with studentachievement. After controlling for individual socio-economic background andcommunity conditions, the most important of these school factors were disciplinaryclimate, student behaviour, student-teacher ratios, teacher support, offering ofextracurricular activities, and teacher specialisation.

1. IntroductionDriven by rapid technological change and the globalisation of markets, the 21st

century affords incredible opportunity to Canada. Policy makers at all levels inCanada are interested in ensuring that all Canadians, including Canadians living inrural areas, have the skills and knowledge to capitalise on this opportunity and theflexibility to adapt to change. Schools in rural areas play a key role in generating therequired skills and knowledge, including the foundation skills that support and enableefficient learning in adulthood.

This study uses data from the Youth in Transition Survey (YITS) and theProgramme for International Assessment (PISA) to measure the difference in readingperformance between students in rural and urban schools in each province and toidentify factors that may help to explain rural-urban differences. The study looks atsome of the factors that help us to understand differences in the reading performanceof rural and urban students and examines how differences in family, school andcommunity environments relate to the rural-urban reading gap.

For this study, a preliminary analysis of rural and urban student populationswas carried out to identify a set of variables that were consistently different for ruraland urban student populations. These variables were then used to develop a modelwhich could test which of the family, school, and community characteristics exertedthe most influence in explaining the rural-urban reading gap. Because the mostconsistent differences between rural and urban students reflect the family andcommunity background of students and are not subject to change in the short term(for example, parent’s education), a second model was then developed to identifyschool characteristics that were not reported at the highest levels in rural schools inorder to identify which of these characteristics has a strong relationship with studentachievement.

7



For the purposes of this analysis, schools were identified as urban or ruralaccording to their location in or their proximity to large urban areas. Urban schoolsare those located in areas identified by Statistics Canada as Census MetropolitanAreas (CMA) and Census Agglomerations (CA). Schools in the remaining ruraland small town areas (RST) were classified as rural.2

What is a rural school?

Rural schools are those located in Rural and small town areas (RST).Rural and Small Town (RST) refers to the population living outside thecommuting zone of Larger Urban Centres (LUCs) – specifically, outsideCensus Metropolitan Areas (CMA) and Census Agglomerations (CA).RST includes all municipalities with urban populations of 1,000 to 9,999and rural areas, where less than 50 percent of the employed individualscommute to the urban core of a CMA/CA.

Urban schools are located in CMAs and CAs and are thus located inurban cores, together with adjacent rural and urban areas that have ahigh degree of economic and social integration with that urban area.A CMA has an urban core of 100,000 or over and includes allneighbouring municipalities where 50 percent or more of the labourforce commutes to the urban core. A CA has an urban core of 10,000to 99,999 and abides by the same commuting rule as a CMA.

This definition was chosen as the best representation of the urban orrural nature of the community to which 15-year-olds are exposed. Sincethis definition is based on actual commuting patterns, it reflects thedegree to which 15-year-olds are likely to have access to an urbancentre and the facilities, educational institutions, and labour marketopportunities that it may provide.

For a full description of the rural definition, see Appendix B.

The Programme for International Student Assessmentand the Youth in Transition Survey

In the spring of 2000, a broad sample of Canadian 15-year-oldsparticipated in the Programme for International Student Assessment(PISA). PISA is a project developed by the Organisation for EconomicCo-operation and Development (OECD) as a way to measure the skilllevels of students in member countries and to understand whatcharacteristics of students and schools influence the level anddistribution of reading, mathematics and science skills among youth.In Canada, the study was carried out in conjunction with the Youth inTransition Survey (YITS), which collected information from studentsand parents about student characteristics and experience. PISA isadministered in Canada through a partnership of the Council ofMinisters of Education, Canada, Human Resources DevelopmentCanada and Statistics Canada.

8



What is PISA?

The Programme for International Student Assessment (PISA) is acollaborative effort among member countries of the Organisation forEconomic Co-operation and Development (OECD) to assess regularlythe performance of 15-year-olds in three domains—reading literacy,mathematical literacy and scientific literacy—through a commoninternational test. Three PISA cycles have been planned, each onefocussing on a different literacy domain. In 2000, the major focus wasreading literacy, with mathematical and scientific literacy as minordomains. As a result, there were fewer mathematics and science itemsincluded in the assessment and these items were administered to asub-sample of participants. Mathematical and scientific literacy willbe the focus in 2003 and 2006, respectively.

Thirty-two countries participated in PISA 2000. In Canada,approximately 30,000 15-year-old students from more than 1,000schools took part. A large Canadian sample was drawn so thatinformation could be provided at both national and provincial levels.

The PISA 2000 survey included a direct assessment of students’skills through reading, mathematics and science tests as well asquestionnaires collecting background information from students andschool principals.

The first results for PISA 2000 are available in the report, Measuringup: the Performance of Canada’s youth in reading, mathematics andscience – OECD PISA Study – First results for Canadians aged 15 (81-590-XIE). This publication is available electronically without charge,through the Internet at www.statcan.ca, www.pisa.gc.ca, andwww.cmec.ca. In addition, the International OECD Report, Knowledgeand Skills for Life: First results from the OECD Programme forInternational Student Assessment, is available on the OECD’s websitewww.pisa.oecd.org.

What is YITS?

The Youth in Transition Survey (YITS) is a new Canadian longitudinalsurvey designed to examine the major transitions in young people’slives, particularly with respect to education, training and work. Surveyresults will provide a deeper understanding of the nature and causesof challenges young people face as they manage their transitions. Thesurvey will help support policy planning and decision making thataddresses these problems.

YITS will examine key transitions in the lives of youth, such as thetransition from high school to postsecondary education and fromschooling to the labour market. The factors that determine high schoolcompletion are examined, as well as the effects of school experienceson educational and occupational outcomes, and the contribution of

9



work experience programs, part-time jobs and volunteer activities. Tocollect this information, current plans are to survey the same group ofyoung people every two years, over a period of several years. Thesecond survey cycle of YITS took place in 2002.

Two different age groups are participating in YITS, the 15-year-oldcohort who also participated in PISA, and an 18- to 20-year-old cohort.The youth aged 18 to 20, who were surveyed in 2000 as part of theYITS project, did not participate in PISA. Results for the 18- to 20-year-old YITS cohort can be found in At a Crossroads: First results for the 18to 20-year-old cohort of the Youth in Transition Survey (81-591-XIE,free) available through the Internet at www.statcan.ca andwww.pisa.gc.ca.

2. Urban students outperform rural studentsin reading

In the PISA 2000 study, Canadian students performed very well by internationalstandards. Canada scored near the top in reading performance. Only students inFinland performed significantly better than Canadian students in reading. Othercountries that performed as well as Canada were New Zealand, Australia, Ireland,and Japan.3

While these results are noteworthy, rural students did not perform quite sowell. At the national level, students from urban schools significantly outperformedstudents from rural schools in reading. In all provinces, except Nova Scotia andManitoba, there were differences in the reading performance of rural and urbanstudents. In only four of these provinces, Newfoundland and Labrador, PrinceEdward Island, New Brunswick and Alberta, were the differences significant.4 It isworth noting, however, that rural students in Alberta still performed well, exceedingthe Canadian national average, and better than urban students in some otherprovinces.

Urban studentssignificantlyoutperformed ruralstudents inNewfoundland andLabrador, PrinceEdward Island, NewBrunswick andAlberta. However,rural students inAlberta stillperformed well,exceeding thenational averageand scoring betterthan urban studentsin some otherprovinces.

What is YITS?

10



Figure 1

Mean reading scores for students in urban and rural schools comparedto provincial averages

Table 1

Student performance in reading for rural and urban schools

Reading Performance

Rural schools Urban schools Provincial average

Newfoundland and Labrador 501 538 517Prince Edward Island 504 526 517Nova Scotia 519 522 521New Brunswick 491 510 501Quebec 527 538 536Ontario 520 535 533Manitoba 527 531 529Saskatchewan 523 533 529Alberta 536 557 550British Columbia 530 539 538Canada 523 538 534

Note: Scores are standardized with an OECD mean of 500, and a standard deviation of 100.Where rural-urban differences in average scores are statistically significant with a 99%level of confidence, scores are bolded.

490

500

510

520

530

540

550

560

Nfld.Lab P.E.I. N.S. N.B. Que. Ont. Man. Sask. Alta. B.C. Canada

URBAN

Provincial average

RURAL

Mean scores

11



0

20

40

60

80

Nfld

.Lab

.

N.B

.

Sas

k.

P.E

.I.

N.S

.

Man

.

Alta

.

Can

ada

Que

.

B.C

.

Ont

.

%

In the international context of the PISA study, students in all provincesperformed well in reading literacy. In fact, all ten provinces performed above theOECD average (of 500). At the provincial level, students in Alberta performedsignificantly better than the Canadian average. Students in Newfoundland andLabrador, Prince Edward Island, Nova Scotia, and New Brunswick, however,performed at levels significantly lower than the Canadian average.

The performance of rural students generally followed the same geographicpattern as the overall provincial averages. However, while the overall provincialresults for Nova Scotia were similar to the other Atlantic Provinces, the rural-urbandifference in performance that appeared in Newfoundland and Labrador, PrinceEdward Island and New Brunswick did not show up in Nova Scotia. In fact, resultsfor Nova Scotian rural students were not significantly lower than for rural studentsin the rest of Canada. Again, Alberta stood out with an average reading performancefor rural students significantly above the national rural average.

The existence of the rural-urban differences in reading performance in someprovinces raises an important question: What are the characteristics of rural students,schools, and communities that can explain these lower results? In what ways arerural students different from urban students and which of these differences liesbehind their lower performance? The first part of this analysis describes how ruraland urban students and their environments differ (Section 3). This leads to a discussionof the characteristics of students, their families, schools and communities that bestexplain the different results for rural and urban students (Section 4). Finally, there isa discussion of school characteristics that were not reported at the highest levels inrural schools in order to identify which of these characteristics has a strong relationshipwith student achievement. (Section 5).

The relative size of rural populations varies by province. While 58%of Newfoundland and Labrador students were in rural schools, onlyabout 15% of students in British Columbia and Ontario were in ruralschools.

Figure 2

Percentage of students attending rural schools by province

12



The relative size of rural populations differs considerably by province. 5

In Newfoundland and Labrador, over half of the 15-year-olds in thePISA study were in rural schools. Considering the low readingperformance of these rural youth, it is clear that the overall provincialresults are strongly influenced by the results of students in the ruralschools. In Alberta, on the other hand, rural students, who also performedat levels significantly lower than their urban counterparts, representedless than one-third of the total student population and thus had a lesserimpact on the overall provincial results.

The provinces with the largest proportion of rural students wereNewfoundland and Labrador, New Brunswick and Saskatchewan.Urban student populations were the most dominant in Ontario, BritishColumbia and Quebec.

3. How do rural and urban students andschools differ?

The first stage in this analysis was an investigation of a variety of student, schooland community characteristics to determine how rural and urban student populationsdiffered and where those differences were consistent with differences in studentreading performance. This information was then used in hypotheses about thedeterminants of the rural-urban reading gap.

Individual student behaviourA variety of factors were available to describe student behaviours or the nature ofstudents’ relationships with others: reading behaviours, social communication withparents, student behaviour and discipline in the classroom (student group behaviours),student-teacher relationships and support from teachers. (See Appendix A forprovincial data tables).

Generally, there were no systematic rural-urban differences in the variablesthat describe personal behaviours and relationships such as reading behaviours andsocial interaction with parents. Enjoyment of reading, for example, which showeda strong correlation with reading performance in the PISA study, was the same forrural and urban students in most provinces with two notable exceptions. InNewfoundland and Labrador and Alberta, rural students reported levels of readingenjoyment significantly lower than those of urban students.

For the most part, rural and urban students reported the same levels of socialinteraction with their parents. In addition, there was generally no difference in ruraland urban student reports on the disciplinary environment of the school or the levelof teacher support or student-teacher relations.

In terms ofindividual student

behaviour andrelationships with

parents andteachers there were

generally nodifferences between

rural and urbanstudents.

13



Family backgroundIn all provinces, the parents of rural students had jobs with lower occupationalstatus, on average, than did the parents of urban students6. The parents of ruralstudents also had significantly lower levels of educational attainment than the parentsof urban students, except in Ontario, Alberta, and British Columbia.

The first results of PISA 2000 showed that the number of books in the homeis an important indicator of reading performance possibly as a reflection of a homeenvironment that encourages reading. In Newfoundland and Labrador, NewBrunswick and Quebec, rural students reported significantly fewer books at homethan did urban students, but elsewhere there was no difference in the proportion ofstudents from homes with more books.

The nature of interactions with parents was also an important factor in readingperformance. Students’ reading performance generally benefited from parents withwhom they could discuss books, television shows, and political or social issues(parental academic interest). This is noteworthy, as urban students reported higherlevels of this kind of interaction with parents in Newfoundland and Labrador, PrinceEdward Island, Quebec, Manitoba, and Alberta.

Urban students in most provinces reported significantly higher levels of homecultural possessions (such as classical literature, books of poetry and works of art)and educational resources (such as a dictionary, a quiet place to study, a desk,textbooks, and calculators). The exceptions were Nova Scotia and British Columbia.In Ontario, there was no significant difference in the presence of cultural possessionsin the home and in Prince Edward Island there was no difference in levels of homeeducational resources.

Rural students were less likely to participate in cultural activities such as goingto museums and attending concerts, likely because of access to cultural facilities. Inmost provinces, however, rural students were just as likely as urban students toparticipate in extracurricular activities at school and outside of school.7

TransportationOne variable of particular interest in this analysis because it is often cited in discussionsabout rural schooling was the amount of time students spent getting to school. Ruralstudents in most provinces reported spending more time travelling to school thanurban students. The only exceptions were Newfoundland and Labrador, Quebecand Alberta where there was no significant difference in the proportion of studentswith long commutes. The likelihood of longer transportation times varied by provincefor both rural and urban students. About one-quarter of rural students in most provincesreported a commute of 30 minutes or more. The only exceptions were Newfoundlandand Labrador, where only 7% of rural students reported transportation times of thatlength, and Prince Edward Island, where fully 42% of students commuted 30 minutesor longer.

Fewer urban students reported such long commutes, but there was an evengreater range by province. In Newfoundland and Labrador, Manitoba and BritishColumbia, fewer than 10% of urban students had commutes longer than 30 minutes.At the other extreme were urban students in Prince Edward Island and Alberta

Rural students weremore likely to comefrom lower socio-economicbackgrounds.

Rural studentstended to comefrom homes withfewer culturalpossessions andeducationalresources and theywere less likely todiscuss cultural,political or socialissues with theirparents.

Rural students spentsignificantly longergetting to school inmost provinces.

14



0

20

40

60

80

100

Nfld

.Lab

.

P.E

.I.

N.S

.

N.B

.

Que

.

Ont

.

Man

.

Sas

k.

Alta

.

B.C

.

Can

ada

Urban Rural

%

where about one-quarter reported the longer commute times. 8 Overall, there wasno consistent pattern between the differences in average transportation times andaverage reading performance for rural and urban jurisdictions.

Computer and Internet UseRural students were less likely to have a computer or Internet access at home thanurban students, but in most provinces the difference was small. Students in ruralschools in Newfoundland and Labrador, New Brunswick and Quebec were theleast likely to have a computer at home. The rural-urban difference was even higherin terms of access to the Internet. In Newfoundland and Labrador and Quebec, onehalf or fewer of rural students had Internet access at home. In contrast, the percentageof urban students with home Internet access ranged from 62% in Quebec to 79% inAlberta. The only provinces without large rural-urban difference in home Internetaccess were Nova Scotia and Ontario.

Figure 3

Rural students are less likely than urban students to have a computer at home

While ruralstudents were less

likely to haveaccess to

computers and theInternet at home

than urbanstudents. . .

Figure 4

Rural students are even less likely to have access to the Internet at home

0

20

40

60

80

100

Nfld

.Lab

.

P.E

.I.

N.S

.

N.B

.

Que

.

Ont

.

Man

.

Sas

k.

Alta

.

B.C

.

Can

ada

Urban Rural

%

15



While rural students were less likely to have access to computers and theInternet at home, they actually made greater use of computers at school. A higherproportion of rural students than urban students used computers more than once amonth at school. In Nova Scotia, Ontario and Quebec, this difference was notstatistically significant. In Prince Edward Island, New Brunswick, Manitoba,Saskatchewan, and Alberta, rural students used the Internet at school significantlymore often than urban students did.

Education and career aspirationsOver 90% of Canadian 15-year-olds reported that they hope to get a postsecondaryeducation. This is very high when one considers that, according to YITS, only 62%of 18 to 20-year-olds, who were no longer in high school in 2000, had gone on topostsecondary education (McMullen, Bowlby, 2002). Fully three-quarters of the15-year-olds who would like to get a postsecondary education said they would liketo get a university degree. To put this in context, of the 18- to 20-year-olds who hadgone on to postsecondary institutions, only one-third actually attended a universityin their first year.9

The percentage of students who aspire to a postsecondary education wasvery high for both rural and urban students in all provinces (85% or more). Only inPrince Edward Island, Quebec and Manitoba were urban students significantlymore likely to want to get a postsecondary education than rural students.

Rural-urban differences were more noticeable when it came to the decisionto aim for a university rather than a college education. While the largest differencesin university intentions were in Alberta and Quebec, only in Prince Edward Islandwere postsecondary preferences the same for both rural and urban students. InQuebec, there were particularly low rates of university intentions for both rural andurban students. This finding may be a reflection of the very different postsecondaryeducation system that exists in Quebec where CEGEPs, the provincial system ofcommunity colleges which also provide a system of university preparation, play avery different role than colleges in the rest of Canada.

The rural-urban difference in the type of postsecondary education expectedwas also reflected in the career aspirations of students as measured by the occupationalstatus of the job they expect to have when they are thirty years old. In all provinces,urban students had significantly higher career aspirations than rural students. Aswith educational aspirations, the career hopes of students in the Atlantic Provinceswere not significantly different from those of other 15-year-olds across the country.

School characteristicsAs part of the PISA assessment, principals were asked to report on the qualificationsof teachers and to report on the extent to which teacher shortages and the adequacyof material and instructional resources hindered student learning. In terms of thepercentage of mathematics, science and language arts teachers who have universitylevel qualifications in their subject matter, in Newfoundland and Labrador, Manitoba,Saskatchewan and Alberta, urban school principals reported significantly higher

they spent moretime using acomputer or theInternet at schoolthan urban studentsdid.

The vast majority of15-year-olds wouldlike to get somekind ofpostsecondaryeducation.

Rural students in allprovinces also hadsignificantly lowercareer expectationsthan urbanstudents.

While moststudents, both ruraland urban, aspire toa universityeducation, the rateis significantlylower for ruralstudents.

Rural and urbanschools are actuallymuch the samewhen it comes toresources andlearningenvironments.

16



levels of teacher specialisation than rural principals. Only in Prince Edward Islandwere rural principals more likely to report that teacher shortages hindered studentlearning.

This analysis is restricted to the schools that participated in the PISAstudy and the communities in which PISA participants went to school.The information on schools presented in this section was collectedfrom questionnaires completed by the principals in the PISA sample ofschools, that is, in schools attended by 15-year-olds. Communityinformation was gathered from the Census for the communities wherethese schools were located. Because the PISA sample was developedto be representative of the population of 15-year-olds, the schoolinformation cannot be interpreted as representative of all schools, orall high schools, urban or rural. Nor is it representative of all rural andurban communities. This analysis describes, rather, the schools attendedby the students in the study and their communities, and thesecharacteristics are included primarily as possible factors influencingstudent performance, not as characteristics of urban or rural schoolsand communities overall.10

For the most part, there were no reported between the adequacy of resourcesin rural and urban schools. The exceptions were Quebec where principals weremore likely to report that the school buildings were inadequate in urban schools,and Saskatchewan where principals reported that instructional resources were lessadequate in rural schools than in urban schools.

Principals were also asked their perceptions of teacher morale and commitmentand the degree to which they thought that negative teacher behaviour affected studentlearning. Although there was a great deal of variation in these measures betweenprovinces, there was no clear trend with respect to urban–rural differences. Only inSaskatchewan and British Columbia did rural and urban principals report differentmeasures of teacher morale and commitment. In Saskatchewan it was urban principalswho reported the highest levels of teacher morale and commitment, and in BritishColumbia it was the rural principals.

Overall, while there were some differences between rural and urban schools,these were not consistent with the rural-urban reading gaps. For example, reports ofthe proportion of teachers working in their area of specialisation differed significantlybetween rural and urban schools in some provinces with the large reading gaps andsome with small or no reading gaps.

Community characteristicsA variety of community characteristics were included in this analysis. Thesecommunity variables were taken from the 1996 Census and PISA 2000. Althoughinformation from the 1996 Census does not reflect the conditions in the communityat the time of the PISA assessment, it does provide an indication of the communitythat these students have likely been exposed to during their schooling. The analysisassumes that the community has not changed significantly in the four years between

Where there weredifferences between

rural and urbanschools, these were

generally notconsistent with the

rural-urban readinggaps.

While there were noconsistent

differences in thenature of schools in

rural and urbancommunities, there

were big differencesin the

characteristics ofthe communities

themselves.

17



the 1996 Census and PISA 2000, and that the students have been exposed to thatcommunity in the interim. It also assumes that the aggregate characteristics of theparents of the 15-year-olds in the school are indicative of the community environmentof the students.

This information describes the communities where the schools in the PISAstudy were located. This part of the analysis examined unemployment andemployment rates, the percentage of white-collar workers in the community,education levels, individual and family income, and the percentage of young adultsenrolled in postsecondary education.

As one would expect, there was a wide range in these community-levelvariables both between rural and urban areas and across the country. In all fourAtlantic Provinces, the communities where rural schools were located hadsignificantly higher rates of adult unemployment than the urban communities in thestudy. Elsewhere, the difference in unemployment rates in the PISA communitieswas not as pronounced. In Manitoba, the communities where rural schools werelocated actually had lower unemployment rates than communities with urbanschools.

There were more notable differences in employment rates, which indicate thepercentage of adults with jobs and therefore account not only for differences in thenumber of adults looking for work (unemployed), but also those who are notparticipating in the labour force at all. In all provinces except Prince Edward Island,Alberta, and British Columbia, adults in rural communities in the study were lesslikely to have a job than those in urban communities.

The differences in economic conditions of rural and urban communities werealso indicated by the average individual and family income of the communities inthe study. In most provinces, average individual income and average family incomeswere higher in the communities where urban schools were located.11

Adult populations in the areas where urban schools were located had higherlevels of educational attainment in all provinces. The urban communities had higherproportions of adults with both any postsecondary education and specificallyuniversity education. These rates were also reflected in the proportion of jobs in thecommunities that typically require university education. Between 40% and 50% ofthe workforce in the urban communities were in these white-collar jobs. In allprovinces, by contrast, less than 40% of jobs in the rural communities requireduniversity training. Only in Quebec was there no rural-urban difference in this white-collar employment rate.

The last community-level indicator included in this analysis was the percentageof young adults in the community who were enrolled in postsecondary education.This does not necessarily indicate the presence of a postsecondary institution in thecommunity as the Census includes many students in the family home even whenthey are away at school. It does, however, provide some indication of the nature ofpostsecondary participation for youth in the community. In all four Atlantic Provincesand Saskatchewan, the rural communities in the study had the same proportions ofyouth enrolled in postsecondary institutions as the urban communities. PrinceEdward Island, Manitoba and Alberta had the lowest rates of youth postsecondaryparticipation in rural communities, and Quebec, Nova Scotia, Newfoundland andLabrador, and Ontario had the highest.

Adults in ruralcommunities hadless education andfewer of them hadjobs requiring auniversity degree.

Rural areas hadhigherunemploymentrates.

18



4. What rural characteristics lie behind thedifference in reading performance?

The first temptation when comparing student performance from different regions isto look to the schools to find out what it is about the education system that canexplain differing results. The initial analysis of rural-urban differences in this reportdemonstrated, however, that the data collected in PISA show that while there aresome differences between rural and urban schools, they do not reflect the rural-urban reading differences.

There were, however, a number of variables that differed significantly betweenrural and urban communities. An examination of these variables showed that, infact, the rural-urban difference in reading performance that exists in some provincesis best explained by differences in the kinds of jobs in the communities where theseschools were located, specifically in the education-level required by those jobs.

In order to determine which factors best explain the difference between thereading performance of rural and urban students, the actual reading averages ofrural students in each province were compared with the averages they would haveattained if there were no rural-urban difference in a variety of important individual,family, school and community factors. For example, after estimating the relationshipbetween parental education and individual performance, it is possible to calculatethe expected individual performance corresponding to various levels of parentaleducation. Using this technique, this analysis calculates what the averageperformance of students in rural areas would be if their families, schools, andcommunities had the average characteristics of the urban areas in their provinces.

The first phase in this analysis looked at the relationship between a combinationof important individual and family factors and the rural-urban gap. The results,shown in Table B1, indicate that while individual and family characteristics explainsome of the rural-urban difference in student reading performance, their explanatoryvalue is small. When the differences in individual and family factors are taken intoaccount, there remains a large, systematic difference between rural and urban studentsthat still needs to be explained. That is, controlling for the family background ofindividual students, rural students still show lower reading scores.

Column 1 shows the size of the rural-urban gap between the actual readingperformance of students in rural and urban schools (a negative number means thatrural student performance is lower by the number indicated). Column 2 shows thesize of the difference that remains after controlling for individual students’ socio-economic status – that is, the difference between actual reading performance ofurban students and the expected performance of rural students if they had the samefamily socio-economic characteristics as their urban counterparts (parental educationand occupational status at the individual level).

In the provinces where there is a substantial difference in the performance ofstudents in rural and urban schools, only a small part of this difference is explainedby the differences in socio-economic status at the student level. This means that onecannot attribute the rural-urban difference in performance solely to differences inthe socio-economic status of individual students. Urban students would still beexpected to perform better than rural students even if their parents had the same

The difference insome provinces

between theperformance of

students in ruraland urban schools

is not related todifferences in the

schools (which arefew) but rather todifferences in the

nature of workfound in rural and

urbancommunities.

While individualand family

characteristicsexplain some of the

rural-urbandifference in

student readingperformance, their

explanatory value issmall.

19



education and similar jobs. In spite of the strong effect of a student’s family socio-economic status on reading performance found in the first results of PISA, there areother important factors at play in explaining the difference in rural and urban results.

Measures of family and community background

This analysis examined a variety of variables related to familybackground (socio-economic status) of students, as well as thecharacteristics of the communities where their schools were located.Variables describing the socio-economic background of students’families were the highest occupational status and the highest level ofeducation of students’ parents. The occupational status of parents wasderived using the International Socio-economic Index of OccupationalStatus (see Appendix B for definition).

As measures of the community environment of students, the analysisexamined average occupational and educational status of the parentsof 15-year-olds in the school, average income, employment rates,unemployment rates, the proportion of adults with postsecondaryeducation, and specifically with university education, and theproportion of adults working in jobs that generally require universityqualifications (white collar jobs).

The second phase in this analysis looked at the relationship between a varietyof community factors and the remaining rural-urban gap. To do this, a model wasdeveloped that looked at the individual and combined effect of a variety of importantcommunity variables such as community employment rates, the average educationalattainment of the adult population, and the average education and occupationalstatus of all the parents of the 15-year-olds in the school. This model confirmed thatdifferences between rural and urban communities best explain the differencesin rural and urban reading performance. More specifically, community characteristicsrelated to adult education levels have the strongest relationship with the rural-urban gap.

Column 3 of Table B1 shows the remaining difference after controlling forthese community variables as well as individual socio-economic status (parentaloccupation and educational attainment). In all provinces, except Alberta, the rural-urban difference disappeared completely once community variables related to adulteducational attainment were controlled for. In fact, after controlling for the communitylevel variables, rural students were actually expected to outperform students fromurban schools. This means that if the rural communities had these same characteristicsas urban communities in the same province, rural students would be expected toscore higher than urban students.

In all provinces, the relative impact of these variables is considerable. A comparisonof the difference that remains after controlling for individual socio-economic statusand the difference after further controlling for the community variables reveals thatthe community factors far outweigh individual student socio-economic status inaccounting for the difference in rural and urban student performance. For example,

Differencesbetween rural andurban communitiesbest explain thedifferences in ruraland urban readingperformance.

20



in Alberta, where there is an actual difference of 21 points between rural and urbanstudents, when individual socio-economic status is controlled for, there is still adifference of 17 points. However, when the combination of community variables istaken into account, this difference shrinks to only 5 points.

Figure 5

Controlling for the difference in family and community background,students in rural schools would actually outperform their urban counterpartsin most provinces

Note: Predicted rural achievement is the expected average performance of students in ruralschools when we control for family background as well as a combination of communityvariables related to average employment rates, occupational status, and educationalattainment. For a comparison of the effects of individual SES and community factors seeTable B1.

490

500

510

520

530

540

550

560

Nfld.lab. P.E.I. N.S. N.B. Que. Ont. Man. Sask. Alta. B.C.

Urban

Rural

Predicted rural

Average reading scores (OECD average=500)

An examination of the variables that contribute to this model reveals that theaverage occupational status of the parents of the 15-year-olds in the school is thefactor with the strongest relationship with the rural-urban difference in readingperformance. Columns 4 to 10 show the size of the differences after controlling forthe impact of the community factors one at a time. In all provinces, when justaggregate parental occupational status is controlled for, rural students perform aboutas well as or better than urban students. Other important community characteristicsthat contribute to the rural-urban gap are the average educational attainment of adults,the proportion of adults with postsecondary education (and more specificallyuniversity education), and the proportion of workers whose jobs require universitytraining. Community average employment rates and average income had smaller orlarger effects depending on the province. In general, these variables are all relatedto the education-level of jobs in the community. While it is unlikely that any ofthese variables directly causes lower student achievement, they are indicators of theenvironment in which these students learn and look for support for their learning. In

The averageoccupational statusof the parents of the

15-year-olds in theschool is the factor

with the strongestrelationship with

the rural-urbandifference in

readingperformance.

21



general, they may be a visible measure of the degree to which higher educationpays off in the community.

Further analysis is required for a full understanding of the relationship betweenthese community characteristics and student performance. To what extent doeducation levels of adults in the community reflect the availability of role modelswho demonstrate the value of education? To what extent do they reflect theavailability of career options that require further education? These questions andothers have yet to be answered. This study was intended as a means of investigatingthe rural-urban difference with a variety of variables available from YITS/PISA.Other variables available both from the YITS/PISA database can provide moredetailed information about how the community level variables interact with studentseducation and career expectations, family background and parental expectations,as well as peer influence. Furthermore, information collected in 2002 when thesestudents were 17 years old will provide an additional set of information on factorsinfluencing decisions about education and labour market choices.

5. Schools matterThis analysis found that the difference between rural and urban reading performanceis most strongly related to community differences in adult education levels and thenature of work in rural communities. Reducing the difference in rural-urban readingperformance is a long-term project that requires a full understanding of the relationshipbetween community education levels and student performance. The rural-urbangap is not related to differences in rural and urban schools because, for the mostpart, rural and urban schools are much the same

Although there are few differences between rural and urban schools, not allschool variables were reported at maximum levels by either rural or urban principals.This analysis concludes by identifying which of the school characteristics not reportedat maximum levels in rural schools have a strong relationship with studentachievement after controlling for individual socio-economic background andcommunity factors.

School characteristics related to school size and structure, school resources,teacher qualifications and training, and school discipline were included. Valuesassociated with high student performance for each of these variables were specifiedbased on the observed relationships with student performance in PISA 2000. Forexample, the percentage of teachers with university qualifications in their instructionalarea was set at 100%. For variables derived as indices (such as student behaviourand teacher morale and commitment) the value was equivalent to there being noreports of problems for the questions which make up the indices. (See Appendix Cfor a description of the values used).

The analysis then estimated how well students in rural schools in each provincewould be expected to perform if all of the school principals had reported the maximumvalue for the questions that make up each of the selected variables. For example,what would the average reading performance of rural students have been if none ofthe principals had reported that poor school resources were hindering studentlearning?

There are a numberof school variableswhich were notreported at thehighest levels andwhich wereassociated withhigh studentachievement.

22



It should be noted that the use of these variables in the model does not implythat these are the critical school factors governing student performance. There arelikely many other factors that are already reported at high levels that may have agreater impact on student performance if the reported values were reduced.Furthermore, the are a great many school factors for which no information wasavailable from the YITS/PISA survey.

There were a number of specific factors which, when modelled at theirmaximum values, were associated with higher average reading performance forrural students (Table C1).12 The most important of these school factors, aftercontrolling for individual socio-economic background and community conditions,were disciplinary climate, student behaviour, student-teacher ratio, teacher support,offering of extracurricular activities, and teacher specialisation.

When the effects of a variety of combinations of school variables wereexamined, the model showed that a specific combination of four school characteristicswas associated with the highest student performance: disciplinary climate, studentbehaviour, student teacher ratios, and availability of school activities. It is importantto note that because changes in school factors would affect both rural and urbanstudents, they would not likely reduce the rural-urban gap.

Other school characteristics were examined in this analysis, includingachievement press, teacher professional development, total hours of instruction,negative teacher behaviour, school autonomy, teacher participation in decisionmaking, teacher morale, and computer availability. However, the effects of thesevariables on performance were either inconsistent or inadequately described by thestatistical model in this analysis.

After controlling forindividual socio-

economicbackground and

communityconditions, the

most important ofthese school factors

were disciplinaryclimate, student

behaviour, student-teacher ratio,

teacher support,offering of

extracurricularactivities, and

teacherspecialisation.

23



ConclusionsThis analysis of the reading performance of students in rural and urban schoolsshowed that the rural students did not perform as well as their urban counterpartsnationally, and particularly in Newfoundland and Labrador, Prince Edward Island,New Brunswick and Alberta. These rural-urban differences are primarily related todifferences between rural and urban communities, in particular, the averageeducational attainment of adults in the community, community employment rates,the educational requirements and earning capacity of jobs in the community.

How the community context actually has an impact on student performanceis a question for further analysis. For the moment, one can only theorise. It may bethat students do better when there is a critical mass of highly educated role modelsin the community who demonstrate the value of learning. Or it may be simply thatstudents apply themselves as hard as they feel they need to – they learn as much asthey feel they need to learn. If there are fewer jobs in their community that requirehigher skills and higher levels of education, then there may no clear incentive forstudents to push themselves that much harder. More work needs to be done tounderstand the impact of community characteristics, family background, and studentseducational and career aspirations on academic performance. Moreover, this analysiscan be extended to an investigation of how these community or neighbourhoodeffects influence student performance in all communities, including smaller cities,which, like rural areas, may have less educated work forces and limited careeropportunities.

Even after considering the effects of community characteristics, there are manyschool characteristics that strongly influence student performance. There are anumber of school variables which were not reported at the highest levels and whichwere associated with high student achievement. After controlling for individualsocio-economic background and community conditions, the most important of theseschool factors were disciplinary climate, student behaviour, student-teacher ratios,teacher support, offering of extracurricular activities, and teacher specialisation. Thisanalysis did not include an exhaustive set of school and community characteristicsthat might influence student performance, nor did the analysis identify which schoolfactors matter the most to educational outcomes. Further research is required inorder to provide an understanding of how community factors influence studentperformance and exactly how certain school characteristics tend to mitigate theeffect of these community factors.

24



Table A1a

Newfoundland and Labrador

Characteristics of students in rural and urban schools

Rural schools Urban schools Provincial Average

Average (s.e.) Average (s.e.) Average (s.e.)

Individual characteristicsReading achievement 501 3.3 538 5.0 517 2.8Mathematics achievement 500 3.3 522 5.2 509 3.0Science achievement 504 4.0 534 5.5 516 3.4Student cultural activities -0.28 0.03 0.06 0.05 -0.14 0.02Reading enjoyment -0.15 0.03 0.04 0.05 -0.07 0.03Homework - students doing 4 or more hours per week (%) 41.0 1.7 45.0 2.3 43.0 1.4Career expectations 60.2 0.6 64.6 0.9 62.0 0.6Percentage of students who expect to get postsecondary education (%) 90.0 1.1 93.0 1.1 91.0 0.7Percentage of students who expect to get university rather than college (%) 76.0 1.3 86.0 1.6 81.0 1.0Percentage of student with 30 minutes or more transportation time (%) 7.0 0.8 7.0 1.1 7.0 0.6Students participating in school extracurricular activities (%) 61.0 1.5 49.0 2.2 56.0 1.3Students participating in non-school extracurricular activities (%) 57.0 1.6 62.0 1.9 59.0 1.1Using computer at school at least several times per month (%) 63.0 1.6 55.0 2.4 60.0 1.3Using internet at school at least several times per month (%) 58.0 1.7 54.0 2.4 56.0 1.4

Family characteristicsFamily socio-economic status 43.6 0.4 53.1 0.7 47.6 0.4Highest parent educational attainment 5.2 0.0 5.6 0.0 5.4 0.0Students with 100 or more books at home (%) 78.0 1.1 84.0 1.7 81.0 0.9Parental academic interest -0.16 0.03 0.08 0.04 -0.06 0.02Parental social interest -0.04 0.03 -0.03 0.04 -0.04 0.03Home cultural possessions -0.16 0.03 0.15 0.05 -0.03 0.03Home educational resources 0.12 0.03 0.27 0.03 0.18 0.02Computer at home (%) 69.9 1.1 85.2 1.7 76.2 1.0Internet access at home (%) 46.7 1.5 67.1 2.1 55.1 1.3

Student’s school experienceDisciplinary climate 0.15 0.03 0.12 0.04 0.14 0.02Teacher-student relations 0.41 0.04 0.52 0.04 0.45 0.03Teacher support 0.42 0.04 0.45 0.04 0.43 0.03

Note: (s.e.) is the standard error of the average. Items in bold indicate significant differences in average values of variables forurban and rural schools.

APPENDIX A: Tables

25



Table A1b

Newfoundland and Labrador

Characteristics of rural and urban schools and communities



School characteristicsSchools offering extracurricular activities (%) 93.04 (0.73) 95.87 (0.97) 93.81 (0.61)Number of 15-year-olds 78.34 (5.70) 157.75 (12.37) 100.00 (6.40)K to 12 schools (%) 21.16 (4.85) 0.00 (0.00) 15.39 (3.64)Negative teacher behaviours -0.14 (0.09) -0.32 (0.17) -0.19 (0.08)Negative student behaviours 0.38 (0.09) 0.25 (0.12) 0.35 (0.07)Teacher morale -0.01 (0.12) 0.08 (0.23) 0.01 (0.10)Student-teacher ratio 14.96 (0.34) 16.94 (0.45) 15.57 (0.29)Inadequacy of instructional resources 0.55 (0.11) 0.17 (0.17) 0.45 (0.09)Shortage of teachers 0.58 (0.12) 0.64 (0.22) 0.59 (0.11)Professional development (%) 50.96 (5.03) 68.75 (7.36) 55.71 (4.24)Inadequacy of material resources 0.00 (0.12) -0.39 (0.15) -0.11 (0.10)Computer availability 6.14 (0.26) 6.77 (0.46) 6.33 (0.23)Teacher specialization (%) 70.99 (2.65) 87.82 (3.62) 75.48 (2.29)School autonomy -0.55 (0.08) -0.31 (0.10) -0.48 (0.06)Teacher participation to decision making -0.05 (0.13) -0.19 (0.17) -0.09 (0.10)

Community characteristicsPopulation density 97.89 (12.01) 368.82 (77.99) 171.78 (25.78)Adult unemployment rate 29.77 (1.28) 12.54 (0.51) 25.07 (1.22)Adult employment rate 35.79 (1.13) 52.98 (1.00) 40.48 (1.16)White collar employment 33.23 (1.32) 49.13 (1.42) 37.57 (1.26)Post-secondary enrollment 61.10 (1.91) 66.69 (1.37) 62.62 (1.45)Post-secondary education 38.93 (1.08) 55.45 (1.03) 43.43 (1.11)University education 14.13 (0.58) 26.88 (1.33) 17.61 (0.80)Average income ($) 16,727 (542) 22,904 (496) 18,412 (500)Average family income ($) 36,365 (1,204) 50,785 (965) 40,298 (1,119)


26



Table A2a

Prince Edward Island




Individual characteristicsReading achievement 504 3.8 526 3.2 517 2.4Mathematics achievement 508 4.6 514 4.4 512 3.7Science achievement 497 5.0 515 4.0 508 2.7Student cultural activities -0.29 0.03 0.09 0.03 -0.06 0.02Reading enjoyment -0.11 0.05 -0.04 0.04 -0.07 0.03Homework - students doing 4 or more hours per week (%) 21.0 1.5 24.0 1.4 23.0 1.0Career expectations 58.7 0.8 62.8 0.6 61.2 0.5Percentage of students who expect to get postsecondary education (%) 87.0 1.4 92.0 1.0 90.0 0.8Percentage of students who expect to get university rather than college (%) 83.0 1.7 88.0 1.2 86.0 1.0Percentage of student with 30 minutes or more transportation time (%) 42.0 1.7 26.0 1.3 32.0 1.0Students participating in school extracurricular activities (%) 60.0 1.9 64.0 1.6 63.0 1.2Students participating in non-school extracurricular activities (%) 65.0 1.9 71.0 1.6 69.0 1.2Using computer at school at least several times per month (%) 61.0 2.0 42.0 1.7 50.0 1.3Using internet at school at least several times per month (%) 54.0 1.9 37.0 1.6 44.0 1.3

Family characteristicsFamily socio-economic status 45.4 0.7 52.2 0.6 49.5 0.5Highest parent educational attainment 5.3 0.0 5.6 0.0 5.5 0.0Students with 100 or more books at home (%) 74.0 1.6 78.0 1.3 77.0 1.0Parental academic interest -0.19 0.04 -0.05 0.03 -0.11 0.02Parental social interest -0.39 0.04 -0.24 0.03 -0.30 0.02Home cultural possessions -0.46 0.04 -0.23 0.03 -0.32 0.03Home educational resources -0.20 0.05 -0.11 0.04 -0.14 0.03Computer at home (%) 75.9 1.8 84.9 1.3 81.4 1.1Internet access at home (%) 52.2 2.4 63.4 1.5 59.1 1.4



27



Table A2b

Prince Edward Island




School characteristicsSchools offering extracurricular activities (%) 94.04 (3.08) 97.64 (0.59) 95.91 (1.52)Number of 15-year-olds 117.64 (19.05) 188.07 (22.02) 154.16 (15.94)K to 12 schools (%) 4.89 (6.23) 2.73 (4.52) 3.77 (3.73)Negative teacher behaviours -0.14 (0.15) 0.32 (0.24) 0.10 (0.15)Negative student behaviours 0.75 (0.30) 0.66 (0.24) 0.70 (0.19)Teacher morale 1.06 (0.21) 0.57 (0.15) 0.80 (0.13)Student-teacher ratio 18.14 (0.51) 18.61 (0.50) 18.34 (0.35)Inadequacy of instructional resources 0.45 (0.22) -0.15 (0.26) 0.13 (0.18)Shortage of teachers 1.69 (0.35) 0.55 (0.15) 1.09 (0.21)Professional development (%) 97.41 (3.35) 63.88 (10.59) 78.99 (6.80)Inadequacy of material resources 0.05 (0.21) -0.60 (0.15) -0.29 (0.14)Computer availability 5.42 (1.07) 8.55 (0.75) 6.80 (0.76)Teacher specialization (%) 61.70 (5.99) 59.19 (5.59) 60.20 (4.04)School autonomy -0.64 (0.11) -0.37 (0.09) -0.50 (0.07)Teacher participation to decision making -0.46 (0.21) -0.13 (0.12) -0.29 (0.12)

Community characteristicsPopulation density 219.50 (79.47) 553.40 (78.70) 392.63 (63.87)Adult unemployment rate 15.01 (1.10) 10.37 (0.30) 12.60 (0.71)Adult employment rate 56.82 (1.70) 59.41 (0.54) 58.16 (0.88)White collar employment 28.22 (2.71) 43.70 (1.08) 36.25 (2.06)Post-secondary enrollment 52.23 (3.28) 61.30 (1.64) 56.93 (1.97)Post-secondary education 40.12 (2.08) 55.48 (1.43) 48.09 (1.94)University education 16.16 (1.01) 28.63 (1.52) 22.62 (1.52)Average income ($) 18,191 (331) 21,442 (206) 19,876 (370)Average family income ($) 42,749 (1,115) 48,610 (464) 45,788 (814)


28



Table A3a

Nova Scotia




Individual characteristicsReading achievement 519 3.3 522 3.5 521 2.3Mathematics achievement 512 3.7 513 4.3 513 2.8Science achievement 514 5.0 517 4.0 516 3.0Student cultural activities 0.04 0.04 0.18 0.03 0.13 0.02Reading enjoyment 0.11 0.04 0.02 0.04 0.06 0.03Homework - students doing 4 or more hours per week (%) 28.0 1.6 26.0 1.4 27.0 0.9Career expectations 60.2 0.6 64.0 0.6 62.6 0.4Percentage of students who expect to get postsecondary education (%) 90.0 1.0 93.0 0.8 92.0 0.6Percentage of students who expect to get university rather than college (%) 79.0 1.6 88.0 1.0 85.0 0.8Percentage of student with 30 minutes or more transportation time (%) 25.0 1.4 12.0 1.0 17.0 0.8Students participating in school extracurricular activities (%) 60.0 1.8 60.0 1.9 60.0 1.2Students participating in non-school extracurricular activities (%) 66.0 1.5 67.0 1.6 67.0 1.1Using computer at school at least several times per month (%) 49.0 1.6 46.0 1.5 48.0 1.1Using internet at school at least several times per month (%) 47.0 1.5 44.0 1.4 45.0 1.0

Family characteristicsFamily socio-economic status 49.1 0.5 53.3 0.6 51.7 0.4Highest parent educational attainment 5.3 0.0 5.5 0.0 5.4 0.0Students with 100 or more books at home (%) 84.0 1.1 82.0 1.2 83.0 0.8Parental academic interest 0.04 0.03 0.14 0.03 0.10 0.02Parental social interest -0.29 0.04 -0.15 0.03 -0.21 0.02Home cultural possessions -0.16 0.03 -0.06 0.03 -0.10 0.02Home educational resources -0.03 0.04 0.02 0.03 0.00 0.02Computer at home (%) 81.7 1.2 85.3 1.3 83.9 0.9Internet access at home (%) 65.4 1.6 69.1 1.2 67.6 1.0


Note: (s.e.) is the standard error of the average. Items in bold indicate significant differences in average values of variables forurban and rural schools

29



Table A3b

Nova Scotia




School characteristicsSchools offering extracurricular activities (%) 94.76 (1.03) 95.19 (0.70) 94.98 (0.62)Number of 15-year-olds 118.60 (9.51) 170.72 (10.25) 144.89 (7.40)K to 12 schools (%) 6.00 (3.24) 0.59 (1.03) 3.26 (1.70)Negative teacher behaviours -0.17 (0.14) -0.05 (0.10) -0.11 (0.08)Negative student behaviours 0.39 (0.11) 0.38 (0.09) 0.38 (0.07)Teacher morale -0.20 (0.14) 0.03 (0.13) -0.09 (0.10)Student-teacher ratio 16.62 (0.30) 16.61 (0.46) 16.61 (0.28)Inadequacy of instructional resources 0.69 (0.13) 0.44 (0.10) 0.57 (0.08)Shortage of teachers 0.87 (0.16) 0.68 (0.13) 0.77 (0.10)Professional development (%) 82.69 (3.98) 86.10 (3.30) 84.37 (2.58)Inadequacy of material resources 0.36 (0.14) 0.09 (0.10) 0.23 (0.09)Computer availability 5.98 (0.46) 6.36 (0.44) 6.18 (0.32)Teacher specialization (%) 72.05 (3.40) 70.50 (3.41) 71.28 (2.40)School autonomy -0.19 (0.08) -0.19 (0.09) -0.19 (0.06)Teacher participation to decision making -0.26 (0.11) -0.11 (0.10) -0.19 (0.07)

Community characteristicsPopulation density 260.31 (42.30) 485.16 (72.18) 373.75 (43.18)Adult unemployment rate 13.13 (0.83) 9.74 (0.65) 11.42 (0.55)Adult employment rate 48.85 (0.84) 56.26 (1.37) 52.59 (0.88)White collar employment 37.12 (1.18) 46.23 (1.01) 41.71 (0.89)Post-secondary enrollment 62.48 (1.19) 64.54 (0.53) 63.52 (0.65)Post-secondary education 48.53 (1.43) 56.18 (1.19) 52.39 (0.99)University education 20.80 (1.66) 28.59 (1.31) 24.73 (1.11)Average income ($) 19,936 (307) 23,383 (515) 21,675 (342)Average family income ($) 42,661 (768) 50,293 (1,161) 46,512 (785)


30



Table A4a

New Brunswick




Individual characteristicsReading achievement 491 2.5 510 2.6 501 1.8Mathematics achievement 503 3.2 509 3.0 506 2.2Science achievement 489 3.2 504 3.1 497 2.3Student cultural activities -0.11 0.03 0.04 0.03 -0.03 0.02Reading enjoyment -0.02 0.03 0.05 0.03 0.02 0.02Homework - students doing 4 or more hours per week (%) 24.0 1.0 29.0 1.4 27.0 0.9Career expectations 58.5 0.5 63.2 0.5 60.9 0.4Percentage of students who expect to get postsecondary education (%) 87.0 1.0 90.0 0.8 89.0 0.6Percentage of students who expect to get university rather than college (%) 77.0 1.1 84.0 1.0 81.0 0.8Percentage of student with 30 minutes or more transportation time (%) 25.0 1.1 20.0 1.2 22.0 0.8Students participating in school extracurricular activities (%) 52.0 1.2 54.0 1.2 53.0 0.8Students participating in non-school extracurricular activities (%) 57.0 1.3 65.0 1.4 61.0 1.0Using computer at school at least several times per month (%) 45.0 1.2 31.0 1.4 37.0 1.0Using internet at school at least several times per month (%) 41.0 1.3 28.0 1.4 34.0 1.0



Note: (s.e.) is the standard error of the average. Items in bold indicate significant differences in average values of variables forurban and rural schools

31



Table A4b

New Brunswick




School characteristicsSchools offering extracurricular activities (%) 89.23 (1.00) 92.52 (0.95) 90.36 (0.76)Number of 15-year-olds 143.63 (10.81) 286.27 (22.58) 192.60 (13.34)K to 12 schools (%) 9.24 (4.42) 1.72 (2.77) 6.66 (3.07)Negative teacher behaviours 0.30 (0.14) 0.18 (0.13) 0.26 (0.10)Negative student behaviours 0.52 (0.08) 0.82 (0.11) 0.62 (0.07)Teacher morale -0.10 (0.18) -0.05 (0.19) -0.09 (0.13)Student-teacher ratio 16.99 (0.26) 18.30 (0.28) 17.46 (0.21)Inadequacy of instructional resources 0.14 (0.10) -0.06 (0.23) 0.07 (0.10)Shortage of teachers 0.52 (0.14) 0.63 (0.19) 0.56 (0.11)Professional development (%) 59.53 (5.63) 76.05 (6.45) 65.36 (4.38)Inadequacy of material resources -0.20 (0.13) -0.29 (0.20) -0.23 (0.11)Computer availability 5.97 (0.39) 6.83 (0.66) 6.27 (0.35)Teacher specialization (%) 65.15 (4.09) 74.16 (5.44) 68.25 (3.29)School autonomy -0.77 (0.12) -0.58 (0.13) -0.71 (0.09)Teacher participation to decision making -0.18 (0.13) -0.59 (0.14) -0.32 (0.10)

Community characteristicsPopulation density 222.23 (26.00) 263.10 (29.36) 236.26 (19.84)Adult unemployment rate 14.61 (1.21) 8.89 (0.69) 12.65 (0.89)Adult employment rate 50.84 (1.21) 57.11 (1.20) 52.99 (0.97)White collar employment 39.02 (1.30) 48.49 (2.13) 42.27 (1.24)Post-secondary enrollment 58.50 (1.50) 62.63 (2.15) 59.92 (1.24)Post-secondary education 43.85 (1.22) 53.27 (2.25) 47.09 (1.23)University education 19.18 (1.02) 28.34 (2.56) 22.32 (1.22)Average income ($) 20,129 (423) 24,780 (1,785) 21,726 (718)Average family income ($) 43,647 (976) 52,880 (3,211) 46,817 (1,371)


32



Table A5a

Quebec




Individual characteristicsReading achievement 527 4.6 538 3.6 536 3.0Mathematics achievement 543 5.6 552 3.3 550 2.7Science achievement 533 5.3 543 4.1 541 3.4Student cultural activities -0.15 0.04 0.00 0.03 -0.03 0.02Reading enjoyment -0.12 0.07 -0.06 0.02 -0.07 0.02Homework - students doing 4 or more hours per week (%) 28.0 3.6 30.0 1.3 30.0 1.1Career expectations 54.8 0.9 60.4 0.5 59.3 0.4Percentage of students who expect to get postsecondary education (%) 91.0 1.0 95.0 0.6 94.0 0.5Percentage of students who expect to get university rather than college (%) 50.0 2.4 64.0 1.4 61.0 1.3Percentage of student with 30 minutes or more transportation time (%) 24.0 2.7 20.0 1.3 21.0 1.2Students participating in school extracurricular activities (%) 50.0 2.9 47.0 1.2 48.0 0.9Students participating in non-school extracurricular activities (%) 49.0 1.6 55.0 1.1 54.0 0.9Using computer at school at least several times per month (%) 45.0 2.9 37.0 1.5 39.0 1.3Using internet at school at least several times per month (%) 41.0 3.2 34.0 1.4 35.0 1.2

Family characteristicsFamily socio-economic status 47.5 0.8 52.5 0.4 51.5 0.4Highest parent educational attainment 5.2 0.1 5.4 0.0 5.4 0.0Students with 100 or more books at home (%) 59.0 1.8 65.0 1.0 64.0 0.9Parental academic interest 0.00 0.05 0.14 0.02 0.12 0.02Parental social interest -0.23 0.05 -0.22 0.02 -0.22 0.02Home cultural possessions -0.61 0.03 -0.35 0.03 -0.40 0.02Home educational resources -0.18 0.04 -0.02 0.02 -0.05 0.02Computer at home (%) 71.6 2.4 81.6 1.0 79.6 0.9Internet access at home (%) 48.5 2.7 61.9 1.0 59.3 1.0

Student’s school experienceDisciplinary climate -0.03 0.06 0.10 0.03 0.08 0.03Teacher-student relations 0.27 0.07 0.28 0.03 0.28 0.02Teacher support 0.33 0.06 0.33 0.02 0.33 0.02


33



Table A5b

Quebec




School characteristicsSchools offering extracurricular activities (%) 90.77 (1.62) 88.43 (0.91) 88.87 (0.80)Number of 15-year-olds 659.80 (31.29) 637.60 (18.11) 641.77 (15.81)K to 12 schools (%) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)Negative teacher behaviours 0.22 (0.17) 0.28 (0.07) 0.27 (0.07)Negative student behaviours 0.40 (0.16) 0.14 (0.08) 0.19 (0.08)Teacher morale 0.09 (0.13) -0.08 (0.08) -0.05 (0.07)Student-teacher ratio 17.15 (0.74) 17.32 (0.29) 17.28 (0.28)Inadequacy of instructional resources -0.52 (0.16) -0.67 (0.08) -0.65 (0.07)Shortage of teachers 0.04 (0.18) -0.16 (0.08) -0.12 (0.08)Professional development (%) 41.98 (6.61) 43.50 (3.16) 43.21 (2.84)Inadequacy of material resources -0.67 (0.10) -0.59 (0.06) -0.61 (0.05)Computer availability 8.16 (0.39) 10.60 (0.78) 10.11 (0.63)Teacher specialization (%) 81.17 (3.72) 86.27 (1.83) 85.22 (1.64)School autonomy -0.24 (0.15) -0.21 (0.06) -0.22 (0.06)Teacher participation to decision making 0.21 (0.15) -0.01 (0.09) 0.03 (0.07)

Community characteristicsPopulation density 231.75 (63.38) 1,939.63 (168.27) 1,618.76 (1,46.63)Adult unemployment rate 10.28 (1.03) 10.22 (0.27) 10.23 (0.29)Adult employment rate 53.90 (1.10) 54.82 (0.54) 54.64 (0.49)White collar employment 36.73 (1.20) 48.86 (0.65) 46.58 (0.68)Post-secondary enrollment 65.00 (2.25) 70.32 (0.61) 69.32 (0.67)Post-secondary education 38.30 (1.25) 49.40 (0.81) 47.31 (0.78)University education 11.98 (0.69) 22.22 (0.90) 20.30 (0.80)Average income ($) 21,573 (485) 23,705 (464) 23,304 (393)Average family income ($) 46,020 (1,004) 50,672 (1,190) 49,798 (994)


34



Table A6a

Ontario





Family characteristicsFamily socio-economic status 48.5 1.1 55.1 0.5 54.2 0.5Highest parent educational attainment 5.5 0.1 5.6 0.0 5.6 0.0Students with 100 or more books at home (%) 83.0 2.0 79.0 1.0 79.0 0.9Parental academic interest -0.02 0.04 0.10 0.02 0.09 0.02Parental social interest -0.20 0.04 -0.19 0.02 -0.19 0.02Home cultural possessions -0.08 0.05 0.04 0.03 0.02 0.03Home educational resources -0.14 0.04 0.08 0.03 0.05 0.02Computer at home (%) 90.0 1.6 93.5 0.6 93.0 0.6Internet access at home (%) 73.3 2.3 78.0 1.0 77.3 0.9



35



Table A6b

Ontario




School characteristicsSchools offering extracurricular activities (%) 94.20 (0.99) 91.61 (1.14) 92.08 (0.95)Number of 15-year-olds 849.03 (44.83) 928.91 (20.67) 914.42 (18.86)K to 12 schools (%) 0.00 (0.00) 0.92 (0.80) 0.75 (0.65)Negative teacher behaviours -0.34 (0.15) -0.38 (0.07) -0.37 (0.07)Negative student behaviours 0.47 (0.12) 0.27 (0.07) 0.31 (0.06)Teacher morale -0.06 (0.15) -0.01 (0.08) -0.01 (0.07)Student-teacher ratio 15.39 (0.45) 16.19 (0.24) 16.04 (0.21)Inadequacy of instructional resources 0.07 (0.16) -0.20 (0.09) -0.15 (0.08)Shortage of teachers 0.16 (0.19) -0.13 (0.08) -0.07 (0.08)Professional development (%) 48.66 (6.34) 50.04 (3.08) 49.78 (2.76)Inadequacy of material resources -0.51 (0.11) -0.36 (0.08) -0.39 (0.07)Computer availability 3.94 (0.16) 5.96 (0.26) 5.56 (0.22)Teacher specialization (%) 76.09 (3.24) 85.47 (1.35) 83.66 (1.28)School autonomy 0.55 (0.16) 0.55 (0.08) 0.55 (0.07)Teacher participation to decision making 0.11 (0.16) -0.04 (0.08) -0.01 (0.07)

Community characteristicsPopulation density 579.67 (68.31) 1,590.90 (1,20.57) 1,407.54 (103.54)Adult unemployment rate 7.19 (0.51) 7.32 (0.16) 7.30 (0.16)Adult employment rate 54.89 (1.24) 61.43 (0.50) 60.25 (0.50)White collar employment 37.48 (0.95) 48.43 (0.61) 46.44 (0.62)Post-secondary enrollment 60.33 (1.51) 69.25 (0.40) 67.63 (0.50)Post-secondary education 41.71 (0.58) 54.02 (0.55) 51.79 (0.58)University education 12.76 (0.42) 25.75 (0.70) 23.39 (0.68)Average income ($) 22,519 (496) 28,401 (318) 27,335 (323)Average family income ($) 49,204 (1,189) 62,487 (822) 60,078 (801)


36



Table A7a

Manitoba




Individual characteristicsReading achievement 527 4.6 531 4.7 529 3.5Mathematics achievement 536 5.3 532 4.9 533 3.7Science achievement 526 4.5 527 4.6 527 3.6Student cultural activities -0.04 0.04 0.18 0.04 0.11 0.03Reading enjoyment -0.07 0.04 -0.05 0.04 -0.06 0.03Homework - students doing 4 or more hours per week (%) 21.0 2.0 27.0 1.7 25.0 1.3Career expectations 56.1 0.8 61.1 0.7 59.4 0.5Percentage of students who expect to get postsecondary education (%) 85.0 1.5 91.0 1.0 89.0 0.9Percentage of students who expect to get university rather than college (%) 74.0 1.5 85.0 1.2 81.0 1.1Percentage of student with 30 minutes or more transportation time (%) 26.0 2.0 8.0 0.8 14.0 0.9Students participating in school extracurricular activities (%) 67.0 2.0 61.0 1.5 63.0 1.4Students participating in non-school extracurricular activities (%) 67.0 2.2 65.0 1.7 66.0 1.4Using computer at school at least several times per month (%) 74.0 2.3 57.0 1.6 63.0 1.4Using internet at school at least several times per month (%) 60.0 3.1 50.0 1.6 53.0 1.8

Family characteristicsFamily socio-economic status 46.9 0.7 52.3 0.7 50.4 0.5Highest parent educational attainment 5.2 0.1 5.4 0.0 5.4 0.0Students with 100 or more books at home (%) 76.0 1.7 77.0 1.5 77.0 1.1Parental academic interest -0.16 0.04 0.00 0.04 -0.06 0.03Parental social interest -0.34 0.04 -0.32 0.03 -0.33 0.03Home cultural possessions -0.35 0.03 -0.14 0.04 -0.21 0.03Home educational resources -0.30 0.04 -0.09 0.04 -0.16 0.03Computer at home (%) 80.9 1.9 87.1 1.2 85.0 1.0Internet access at home (%) 50.6 1.9 68.7 1.6 62.5 1.4



37



Table A7b

Manitoba




School characteristicsSchools offering extracurricular activities (%) 94.51 (0.78) 95.19 (0.68) 94.87 (0.51)Number of 15-year-olds 124.45 (8.92) 194.49 (12.04) 161.46 (8.32)K to 12 schools (%) 30.68 (6.66) 4.07 (2.69) 16.63 (3.67)Negative teacher behaviours 0.00 (0.11) 0.03 (0.10) 0.02 (0.08)Negative student behaviours 0.27 (0.10) 0.21 (0.12) 0.24 (0.08)Teacher morale 0.24 (0.14) 0.09 (0.11) 0.16 (0.09)Student-teacher ratio 16.06 (0.54) 16.78 (0.34) 16.47 (0.30)Inadequacy of instructional resources 0.19 (0.12) 0.10 (0.10) 0.14 (0.08)Shortage of teachers 0.21 (0.14) 0.30 (0.14) 0.26 (0.10)Professional development (%) 83.29 (4.09) 87.32 (3.42) 85.43 (2.63)Inadequacy of material resources -0.31 (0.11) -0.13 (0.12) -0.21 (0.08)Computer availability 4.28 (0.23) 4.60 (0.22) 4.47 (0.16)Teacher specialization (%) 61.38 (3.28) 79.99 (3.14) 70.80 (2.46)School autonomy 0.66 (0.11) 0.52 (0.13) 0.59 (0.09)Teacher participation to decision making 0.37 (0.16) 0.38 (0.13) 0.37 (0.10)

Community characteristicsPopulation density 385.89 (40.25) 1,189.55 (47.74) 810.47 (50.20)Adult unemployment rate 4.40 (0.54) 6.65 (0.09) 5.59 (0.28)Adult employment rate 56.78 (1.23) 61.89 (0.40) 59.48 (0.67)White collar employment 35.39 (1.12) 46.37 (0.33) 41.19 (0.77)Post-secondary enrollment 51.26 (1.88) 59.97 (0.39) 55.86 (1.00)Post-secondary education 37.84 (1.12) 52.60 (0.30) 45.63 (0.90)University education 15.21 (0.64) 27.42 (0.44) 21.66 (0.70)Average income ($) 20,250 (494) 24,017 (168) 22,240 (308)Average family income ($) 44,716 (1,023) 52,991 (337) 49,088 (651)


38



Table A8a

Saskatchewan




Individual characteristicsReading achievement 523 4.6 533 3.2 529 2.7Mathematics achievement 525 4.3 524 3.5 525 2.9Science achievement 520 4.4 523 4.2 522 3.0Student cultural activities -0.07 0.04 0.14 0.03 0.05 0.02Reading enjoyment -0.12 0.04 -0.16 0.03 -0.14 0.03Homework - students doing 4 or more hours per week (%) 15.0 1.6 22.0 1.3 19.0 1.1Career expectations 55.8 0.6 61.2 0.6 58.9 0.5Percentage of students who expect to get postsecondary education (%) 87.0 1.0 89.0 1.0 88.0 0.7Percentage of students who expect to get university rather than college (%) 76.0 1.7 85.0 1.6 81.0 1.1Percentage of student with 30 minutes or more transportation time (%) 23.0 1.6 11.0 1.0 16.0 1.0Students participating in school extracurricular activities (%) 74.0 1.7 63.0 1.9 68.0 1.4Students participating in non-school extracurricular activities (%) 71.0 1.6 72.0 1.3 72.0 1.0Using computer at school at least several times per month (%) 77.0 2.8 49.0 2.2 61.0 1.6Using internet at school at least several times per month (%) 67.0 3.6 43.0 2.2 53.0 1.9



Note (s.e.) is the standard error of the average. Items in bold indicate significant differences in average values of variables forurban and rural schools.

39



Table A8b

Saskatchewan




School characteristicsSchools offering extracurricular activities (%) 93.71 (0.81) 94.84 (0.73) 94.16 (0.57)Number of 15-year-olds 96.50 (4.32) 197.22 (10.56) 136.96 (6.74)K to 12 schools (%) 44.25 (6.04) 3.46 (2.69) 27.66 (4.18)Negative teacher behaviours -0.17 (0.11) -0.19 (0.15) -0.18 (0.09)Negative student behaviours 0.23 (0.09) 0.26 (0.12) 0.24 (0.07)Teacher morale 0.05 (0.11) 0.77 (0.14) 0.34 (0.09)Student-teacher ratio 16.09 (0.30) 18.59 (0.32) 17.12 (0.25)Inadequacy of instructional resources 0.12 (0.10) -0.37 (0.14) -0.08 (0.09)Shortage of teachers 0.24 (0.11) 0.28 (0.13) 0.26 (0.08)Professional development (%) 65.00 (4.11) 76.62 (5.08) 69.67 (3.22)Inadequacy of material resources -0.29 (0.10) -0.31 (0.10) -0.30 (0.07)Computer availability 5.43 (0.22) 7.87 (1.15) 6.42 (0.50)Teacher specialization (%) 52.98 (2.78) 75.46 (3.31) 61.84 (2.36)School autonomy -0.04 (0.07) -0.07 (0.09) -0.05 (0.06)Teacher participation to decision making 0.34 (0.12) 0.39 (0.14) 0.36 (0.09)

Community characteristicsPopulation density 379.86 (26.03) 1,088.23 (75.51) 664.42 (46.79)Adult unemployment rate 5.26 (0.75) 5.93 (0.22) 5.53 (0.46)Adult employment rate 55.78 (1.24) 62.78 (0.72) 58.59 (0.85)White collar employment 32.70 (1.14) 44.24 (0.94) 37.34 (0.94)Post-secondary enrollment 55.58 (2.23) 59.11 (0.89) 57.00 (1.39)Post-secondary education 38.55 (1.12) 51.72 (0.95) 43.84 (0.98)University education 15.26 (0.73) 25.73 (1.08) 19.46 (0.78)Average income ($) 19,359 (778) 24,616 (457) 21,471 (553)Average family income ($) 41,897 (1,711) 53,860 (1,269) 46,702 (1,263)


40



Table A9a

Alberta




Individual characteristicsReading achievement 536 4.3 557 4.2 550 3.3Mathematics achievement 541 5.0 550 4.3 547 3.3Science achievement 532 4.6 553 4.3 546 3.5Student cultural activities 0.10 0.04 0.22 0.03 0.18 0.02Reading enjoyment -0.14 0.05 0.04 0.03 -0.02 0.03Homework - students doing 4 or more hours per week (%) 26.0 2.0 37.0 1.7 33.0 1.3Career expectations 56.6 1.0 63.4 0.5 61.3 0.4Percentage of students who expect to get postsecondary education (%) 89.0 1.1 92.0 0.6 91.0 0.5Percentage of students who expect to get university rather than college (%) 63.0 2.5 81.0 1.3 76.0 1.2Percentage of student with 30 minutes or more transportation time (%) 28.0 2.4 23.0 1.7 24.0 1.4Students participating in school extracurricular activities (%) 61.0 2.1 58.0 1.4 59.0 1.1Students participating in non-school extracurricular activities (%) 70.0 2.2 67.0 1.4 68.0 1.2Using computer at school at least several times per month (%) 72.0 2.7 54.0 1.5 60.0 1.3Using internet at school at least several times per month (%) 67.0 3.0 50.0 1.6 56.0 1.4

Family characteristicsFamily socio-economic status 50.5 0.9 55.6 0.6 54.1 0.5Highest parent educational attainment 5.4 0.0 5.5 0.0 5.5 0.0Students with 100 or more books at home (%) 77.0 1.7 81.0 1.3 79.0 1.0Parental academic interest -0.06 0.04 0.15 0.03 0.09 0.02Parental social interest -0.21 0.04 -0.11 0.02 -0.14 0.02Home cultural possessions -0.21 0.06 0.08 0.03 -0.01 0.03Home educational resources -0.07 0.05 0.15 0.03 0.09 0.03Computer at home (%) 87.2 1.2 91.3 0.8 90.0 0.7Internet access at home (%) 61.3 1.7 79.0 1.3 73.5 1.1



41



Table A9b

Alberta




School characteristicsSchools offering extracurricular activities (%) 92.17 (1.36) 95.83 (0.69) 94.67 (0.65)Number of 15-year-olds 352.53 (16.22) 380.63 (13.85) 371.73 (10.79)K to 12 schools (%) 27.69 (7.50) 2.12 (1.63) 10.19 (2.82)Negative teacher behaviours -0.06 (0.12) -0.19 (0.10) -0.14 (0.08)Negative student behaviours 0.28 (0.12) 0.28 (0.08) 0.28 (0.07)Teacher morale 0.20 (0.14) 0.48 (0.11) 0.39 (0.09)Student-teacher ratio 18.72 (0.55) 19.60 (0.36) 19.32 (0.30)Inadequacy of instructional resources -0.08 (0.14) -0.19 (0.12) -0.16 (0.09)Shortage of teachers 0.40 (0.16) 0.31 (0.13) 0.34 (0.10)Professional development (%) 74.09 (5.06) 72.28 (4.23) 72.87 (3.29)Inadequacy of material resources -0.12 (0.13) 0.05 (0.10) -0.01 (0.08)Computer availability 4.93 (0.23) 5.40 (0.25) 5.25 (0.18)Teacher specialization (%) 58.59 (4.04) 79.34 (2.43) 72.24 (2.31)School autonomy 0.36 (0.11) 0.26 (0.08) 0.29 (0.07)Teacher participation to decision making 0.72 (0.15) 0.39 (0.13) 0.50 (0.10)

Community characteristicsPopulation density 415.20 (40.56) 809.43 (38.89) 684.59 (33.88)Adult unemployment rate 6.30 (0.55) 6.19 (0.13) 6.23 (0.19)Adult employment rate 63.93 (1.65) 67.19 (0.45) 66.16 (0.62)White collar employment 37.55 (0.98) 46.63 (0.51) 43.75 (0.61)Post-secondary enrollment 52.62 (1.62) 60.95 (0.52) 58.31 (0.72)Post-secondary education 46.46 (1.19) 57.87 (0.47) 54.25 (0.69)University education 15.39 (0.62) 26.01 (0.65) 22.65 (0.66)Average income ($) 24,189 (650) 27,383 (371) 26,372 (353)Average family income ($) 51,631 (1,372) 59,471 (834) 56,989 (788)


42



Table A10a

British Columbia





Family characteristicsFamily socio-economic status 49.7 0.7 54.0 0.6 53.3 0.5Highest parent educational attainment 5.5 0.0 5.5 0.0 5.5 0.0Students with 100 or more books at home (%) 85.0 2.1 79.0 1.0 79.0 1.0Parental academic interest 0.00 0.06 0.16 0.03 0.14 0.02Parental social interest -0.21 0.04 -0.20 0.02 -0.20 0.02Home cultural possessions -0.08 0.11 0.03 0.03 0.01 0.03Home educational resources -0.09 0.06 0.07 0.03 0.04 0.02Computer at home (%) 88.6 2.1 93.7 0.8 92.9 0.7Internet access at home (%) 65.8 3.4 78.0 1.3 76.1 1.2



43



Table A10b

British Columbia




School characteristicsSchools offering extracurricular activities (%) 94.52 (1.10) 95.42 (0.44) 95.28 (0.41)Number of 15-year-olds 425.33 (21.66) 396.63 (8.31) 401.07 (7.80)K to 12 schools (%) 0.56 (1.81) 4.10 (1.95) 3.57 (1.69)Negative teacher behaviours -0.33 (0.20) -0.13 (0.09) -0.16 (0.08)Negative student behaviours 0.11 (0.21) 0.20 (0.07) 0.19 (0.07)Teacher morale 0.76 (0.24) 0.15 (0.10) 0.24 (0.10)Student-teacher ratio 17.46 (1.01) 17.38 (0.30) 17.39 (0.30)Inadequacy of instructional resources -0.28 (0.22) -0.24 (0.10) -0.25 (0.09)Shortage of teachers -0.18 (0.20) -0.30 (0.08) -0.28 (0.07)Professional development (%) 99.24 (1.70) 80.16 (3.33) 82.76 (2.95)Inadequacy of material resources -0.32 (0.23) -0.34 (0.08) -0.33 (0.08)Computer availability 5.73 (0.58) 6.20 (0.29) 6.12 (0.26)Teacher specialization (%) 73.67 (4.99) 83.13 (1.82) 81.73 (1.74)School autonomy 0.35 (0.22) 0.36 (0.09) 0.36 (0.08)Teacher participation to decision making 0.07 (0.29) 0.22 (0.10) 0.20 (0.10)

Community characteristicsPopulation density 412.30 (90.03) 1,128.72 (132.13) 1,018.05 (114.88)Adult unemployment rate 8.84 (0.59) 8.08 (0.23) 8.19 (0.22)Adult employment rate 56.77 (2.03) 61.23 (0.44) 60.54 (0.51)White collar employment 37.48 (1.47) 45.19 (0.72) 44.00 (0.69)Post-secondary enrollment 56.08 (1.50) 61.99 (0.70) 61.08 (0.66)Post-secondary education 50.12 (1.63) 56.17 (0.66) 55.23 (0.64)University education 17.82 (1.33) 25.77 (0.84) 24.54 (0.78)Average income ($) 24,808 (874) 27,136 (422) 26,777 (388)Average family income ($) 51,627 (1,890) 58,810 (1,103) 57,700 (1,002)


44



Table A11a

Canada


Rural schools Urban schools Canada


Individual characteristicsReading achievement 523 1.9 538 2.0 534 1.6Mathematics achievement 527 2.1 535 1.8 533 1.4Science achievement 521 2.2 532 2.0 529 1.6Student cultural activities -0.01 0.02 0.16 0.02 0.12 0.01Reading enjoyment -0.06 0.02 0.01 0.01 0.00 0.01Homework - students doing 4 or more hours per week (%) 28.0 1.1 37.0 0.7 35.0 0.6Career expectations 57.2 0.4 62.8 0.2 61.6 0.2Percentage of students who expect to get postsecondary education (%) 90.0 0.6 94.0 0.3 93.0 0.2Percentage of students who expect to get university rather than college (%) 66.0 1.0 77.0 0.7 75.0 0.6Percentage of student with 30 minutes or more transportation time (%) 24.0 0.9 16.0 0.5 17.0 0.5Students participating in school extracurricular activities (%) 60.0 1.0 57.0 0.7 58.0 0.6Students participating in non-school extracurricular activities (%) 62.0 0.9 64.0 0.5 63.0 0.4Using computer at school at least several times per month (%) 59.0 1.2 48.0 0.9 50.0 0.7Using internet at school at least several times per month (%) 52.0 1.5 43.0 1.0 45.0 0.8

Family characteristicsFamily socio-economic status 48.1 0.4 54.1 0.3 52.8 0.2Highest parent educational attainment 5.3 0.0 5.5 0.0 5.5 0.0Students with 100 or more books at home (%) 75.0 0.9 76.0 0.6 76.0 0.5Parental academic interest -0.05 0.02 0.11 0.01 0.08 0.01Parental social interest -0.24 0.02 -0.20 0.01 -0.21 0.01Home cultural possessions -0.28 0.02 -0.07 0.02 -0.12 0.01Home educational resources -0.14 0.02 0.04 0.01 0.00 0.01Computer at home (%) 81.8 0.9 89.6 0.4 87.9 0.3Internet access at home (%) 59.7 1.2 73.1 0.6 70.2 0.5



45



Table A11b

Canada


Rural schools Urban schools Canada


School characteristicsSchools offering extracurricular activities (%) 92.88 (0.47) 92.16 (0.37) 92.32 (0.31)Number of 15-year-olds 466.11 (21.20) 648.21 (11.29) 608.66 (10.21)K to 12 schools (%) 11.31 (2.05) 1.44 (0.41) 3.59 (0.56)Negative teacher behaviours -0.09 (0.06) -0.13 (0.03) -0.12 (0.03)Negative student behaviours 0.36 (0.05) 0.24 (0.03) 0.27 (0.02)Teacher morale 0.11 (0.06) 0.07 (0.03) 0.08 (0.03)Student-teacher ratio 16.69 (0.22) 17.08 (0.11) 16.99 (0.10)Inadequacy of instructional resources -0.03 (0.06) -0.29 (0.04) -0.24 (0.03)Shortage of teachers 0.25 (0.07) -0.05 (0.03) 0.02 (0.03)Professional development (%) 61.83 (2.46) 58.66 (1.33) 59.36 (1.17)Inadequacy of material resources -0.34 (0.05) -0.35 (0.03) -0.35 (0.03)Computer availability 5.64 (0.16) 7.12 (0.19) 6.79 (0.16)Teacher specialization (%) 70.21 (1.55) 83.55 (0.67) 80.46 (0.65)School autonomy 0.11 (0.06) 0.24 (0.03) 0.21 (0.03)Teacher participation to decision making 0.20 (0.06) 0.06 (0.03) 0.09 (0.03)

Community characteristicsPopulation density 372.72 (22.37) 1,440.47 (50.79) 1,208.56 (42.16)Adult unemployment rate 9.59 (0.49) 8.11 (0.10) 8.43 (0.13)Adult employment rate 54.81 (0.64) 60.11 (0.23) 58.96 (0.24)White collar employment 36.60 (0.47) 47.65 (0.25) 45.25 (0.26)Post-secondary enrollment 58.84 (0.79) 66.83 (0.24) 65.10 (0.27)Post-secondary education 42.37 (0.52) 53.54 (0.26) 51.11 (0.27)University education 14.73 (0.37) 25.13 (0.30) 22.87 (0.28)Average income ($) 21,777 (265) 26,490 (162) 25,467 (151)Average family income ($) 46,873 (570) 57,680 (404) 55,333 (365)


46



Table A12

Significant differences in student characteristics for urban and rural schools

Nfld.Lab. P.E.I. N.S. N.B. Que. Ont. Man. Sask. Alta. B.C. Canada

Reading achievement S S . S . . . . S . S

Mathematics achievement S . . . . . . . . . .

Science achievement S . . S . . . . S . S

Student cultural activities S S S S S . S S . S S

Reading enjoyment S . . . . . . . S . S

Homework - students doing 4 or more hours per week (%) . . . S . S . S S S S

Career expectations S S S S S S S S S . S

Percentage of students who expect to get postsecondary education (%) . S . . S . S . . . S

Percentage of students who expect to get university rather than college (%) S . S S S S S S S S S

Percentage of student with 30 minutes or more transportation time (%) . S S S . S S S . S S

Students participating in school extracurricular activities (%) S . . . . . S S . . .

Students participating in non-school extracurricular activities (%) . . . S S . . . . . .

Using computer at school at least several times per month (%) S S . S . . S S S S S

Using internet at school at least several times per month (%) . S . S . . S S S . S

Family characteristicsFamily socio-economic status S S S S S S S S S S S

Highest parent educationalattainment S S S S S . S S . . S

Students with 100 or more books at home (%) S . . S S . . . . . .

Parental academic interest S S . . S . S . S . S

Parental social interest . S S . . . . . . . .

Home cultural possessions S S . S S . S S S . S

Home educational resources S . . S S S S S S . S

Computer at home (%) S S . S S . S . S . S

Internet access at home (%) S S . S S . S . S S S

Student’s school experienceDisciplinary climate . . . . . . . . . . .

Teacher-student relations . . . . . . S . . . .

Teacher support . . . . . . . . . . .

Note: S identifies characteristics where the difference between urban and rural schools is statistically significant with a 99%level of confidence.

47



Table A13

Significant differences in school and community characteristics for urban and rural schools

Nfld.Lab. P.E.I. N.S. N.B. Que. Ont. Man. Sask. Alta. B.C. Canada

Schools offering extracurricular activities (%) . . . . . . . . . . .

Number of 15-year-olds S . S S . . S S . . S

K to 12 schools (%) S . . . . . S S S . S

Negative teacher behaviours . . . . . . . . . . .

Negative student behaviours . . . . . . . . . . .

Teacher morale . . . . . . . S . . .

Student-teacher ratio S . . S . . . S . . .

Inadequacy of instructional resources . . . . . . . S . . S

Shortage of teachers . S . . . . . . . . S

Professional development (%) . S . . . . . . . S .

Inadequacy of material resources . . . . S . . . . . .

Computer availability . . . . . S . . . . S

Teacher specialization (%) S . . . . . S S S . S

School autonomy . . . . . . . . . . .

Teacher participation to decision making . . . . S . . . . . .

Community characteristicsPopulation density S S S . . S S S S S S

Adult unemployment rate S S S S . . S . . . S

Adult employment rate S . S S S S S S . . S

White collar employment S S S S . S S S S S S

Post-secondary enrollment . . . . S S S . S S S

Post-secondary education S S S S S S S S S S S

University education S S S S S S S S S S S

Average income ($) S S S . S S S S S . S

Average family income ($) S S S . . S S S S S S

Note. S identitifies characteristics where the difference between urban and rural schools is statistically significant withapproximately 99% confidence.

48



Table B1

Size of rural-urban differences controlling for individual student socio-economic status and a varietyof community factors

Controlling for individual SES and.....

Controlling Controlling ... averageActual for for parentalrural- individual individual occupa- ... adult ... average ... % of

urban gap socio- SES and tional ... employ- univer- post- jobs ... averagein reading economic community status of ... average ment sity secondary requiring educational

achievement status (SES) factors school income rates attainment attainment university attainment

Newfoundland and Labrador -37 -27.8 2.1 1.0 -20.0 -17.7 -19.1 -11.9 -15.7 -6.2PrinceEdward Island -21.9 -14.5 6.3 6.7 -10.4 -13 -5.9 0.4 -2.6 6.3Nova Scotia -3.3 1.1 14.3 14.1 5.5 5.5 6.5 8.5 8.1 13.7New Brunswick -18.1 -11.4 10.3 10.0 -5.5 -7.7 -5.1 -2.3 -4.2 4.0Quebec -10.5 -5.5 7.4 9.6 -2.8 -5.0 1.6 5.3 3.8 7.7Ontario -14.8 -9.7 5.3 10.7 -2.2 -5.9 -0.7 2.2 -1.3 -5.5Manitoba -3.8 1.4 16.6 17.5 6.2 4.4 9.9 15.8 9.8 14.0Saskatchewan -10.1 -4.4 12.9 15.7 2.3 -0.3 2.8 8.4 4.5 5.3Alberta -20.9 -16.9 -5.4 -1.1 -12.8 -15.0 -9.5 -5.7 -9.9 -12.6British Columbia -9.2 -6.0 2.2 7.1 -3.0 -3.4 -0.5 -0.2 -0.1 -3.7

Table C1

Expected differences in grade 10 rural student performance through maximising school characteristics

Originalpredicted Studentgrade 10 Teacher Disciplinary School teacher Student Teacher Standardized Final average support climate activities ratio behaviour specialization testing model

Alberta 541 9 37 6 4 13 4 0 50Newfoundland and Labrador 504 8 35 5 10 14 3 0 53Prince Edward Island 507 8 36 5 5 15 4 0 49Nova Scotia 530 8 36 4 7 14 3 0 50New Brunswick 496 9 36 8 6 14 3 2 53Quebec 546 8 31 7 7 14 2 1 49Ontario 521 8 36 5 9 14 2 0 53Manitoba 532 9 38 4 8 13 4 0 52Saskatchewan 530 8 35 5 8 13 5 0 50British Columbia 533 8 34 4 7 12 2 0 47

49



APPENDIX B

Definitions and ConceptsNote: Several of the measures in this report reflect indices that summarise responsesfrom students or school representatives (typically principals) to a series of relatedquestions. It is important to note that negative values in an index do not necessarilyimply that students responded negatively to the underlying questions. A negativevalue merely indicates that a group of students (or all students, collectively, in asingle country) or principals responded less positively than all students or principalsdid on average across OECD countries. Likewise, a positive value on an indexindicates that a group of students or principals responded more favourably, or morepositively, than students or principals did, on average, in OECD countries.

I. Measures of reading, mathematics andscience literacy

Reading literacy is defined in PISA as the ability to understand, use and reflect onwritten texts in order to achieve one’s goals, to develop one’s knowledge andpotential, and to participate effectively in society. This definition goes beyond thenotion that reading literacy means decoding written material and literalcomprehension. Reading incorporates understanding and reflecting on texts. Literacyinvolves the ability of individuals to use written information to fulfil their goals, andthe consequent ability of complex modern societies to use written information tofunction effectively.

Mathematical literacy is defined in PISA as the capacity to identify, understandand engage in mathematics, and to make well-founded judgements about the rolethat mathematics plays in an individual’s current and future private life, occupationallife, social life with peers and relatives, and life as a constructive, concerned andreflective citizen. As with reading, the definition revolves around the wider uses ofmathematics in people’s lives rather than being limited to mechanical operations.“Mathematical literacy” is used here to indicate the ability to put mathematicalknowledge and skills to functional use rather than just mastering them within aschool curriculum. To “engage in” mathematics covers not simply physical or socialactions (such as deciding how much change to give someone in a shop) but alsowider uses, including taking a point of view and appreciating things expressedmathematically (such as having an opinion about a government’s spending plans).Mathematical literacy also implies the ability to pose and solve mathematical problemsin a variety of situations, as well as the inclination to do so, which often relies onpersonal traits such as self-confidence and curiosity.

Scientific literacy relates to the ability to think scientifically in a world in whichscience and technology shape lives. Such literacy requires an understanding ofscientific concepts as well as an ability to apply a scientific perspective. PISA definesscientific literacy as the capacity to use scientific knowledge, to identify questions,and to draw evidence-based conclusions in order to understand and help makedecisions about the natural world and the changes made to it through human activity.

50



II. Individual, family, and school characteristicsReading Enjoyment: This index was derived from students’ level of agreementwith the following statements: I read only if I have to; reading is one of my favouritehobbies; I like talking about books with other people; I find it hard to finish books;I feel happy if I receive a book as a present; for me reading is a waste of time; Ienjoy going to a bookstore or a library; I read only to get information that I need;and, I cannot sit still and read for more than a few minutes.

Students’ cultural activities: This index was derived from students’ reports onhow often they had participated in the following activities during the precedingyear: visited a museum or art gallery; attended an opera, ballet or classical symphonyconcert; and watched live theatre.

Homework time: This variable was collected as part of the Youth in TransitionSurvey (YITS). Students were asked about how many hours per week they usuallyspend on homework outside class (during free periods and at home).

Career expectations: Students were asked to report what kind of job they expectto have when they are about thirty years old. This information was then classifiedby occupational status according to the International Socio-Economic Index ofOccupational Status (ISEI) (defined under socio-economic status).

Student’s education expectations: This variable was collected as part of the Youthin Transition Survey (YITS). Students reported what is the highest level of educationthey would like to get.

Transportation time: This variable was collected as part of the Youth in TransitionSurvey (YITS). Students were asked how long it usually takes them to travel toschool one way. This variable included use of all forms of transportation.

Participation in school extracurricular activities: This variable was collected aspart of the Youth in Transition Survey (YITS). Students were asked how many totalhours per week they usually spend participating in school clubs, teams or otherschool-based extracurricular activities.

Participation in non-school extracurricular activities: This variable was collectedas part of the Youth in Transition Survey (YITS). Students were asked how manytotal hours per week they usually spend participating in non-school clubs, teams,lessons or other non-school organised activities.

Computer and Internet at home: Students were asked how many computersthey had in their home and whether they had a link to the Internet in their home.

Computer and Internet use at school: Students were asked how often they usecomputers and the Internet at school.

Parents’ occupational status: Students were asked to report their mothers’ andfathers’ occupation, and to state whether each parent was: in full-time paid work;part-time paid work; not working but looking for a paid job; or “other”. The open-ended responses were then coded in accordance with the International StandardClassification of Occupations (ISCO 1988).

51



The PISA International Socio-Economic Index of Occupational Status (ISEI):was derived from student responses on parental occupation. The index captures theattributes of occupations that convert parents’ education into income. The indexwas derived by the optimal scaling of occupation groups to maximise the indirecteffect of education on income through occupation and to minimise the direct effectof education on income, net of occupation (both effects being net of age). For moreinformation on the methodology, see Ganzeboom, de Graaf and Treiman (1992).The PISA International Socio-Economic Index of Occupational Status is based oneither the father’s or mother’s occupations, whichever is the higher.

Parent’s educational attainment: This variable was derived as the highest levelof attainment achieved by either parent.

Number of books at home: Students reported an estimate of how many booksthere are in their home. They were given a calculation that there are approximately40 books per metre of shelving and were asked not to include magazines.

Parental academic interest: The index of parental academic interest was derivedfrom students’ reports on the frequency with which their parents (or guardians)engaged with them in the following activities: discussing political or social issues;discussing books, films or television programmes; and listening to classical music.

Parental social interest: This index was derived from students’ reports on thefrequency with which their parents (or guardians) engaged with them in the followingactivities: discussing how well they are doing at school; eating the evening mealwith them around a table; and spending time simply talking with them.

Home cultural possessions: This index was derived from students’ reports on theavailability of the following items in their home: classical literature (examples weregiven), books of poetry and works of art (examples were given).

Home educational resources: This index was derived from students’ reports onthe availability and number of the following items in their home: a dictionary, aquiet place to study, a desk for study, textbooks and calculators.

Disciplinary climate: This index summarises students’ reports on the frequencywith which, in their <class of the language of assessment>: the teacher has to waita long time for students to quiet down; students cannot work well; students don’tlisten to what the teacher says; students don’t start working for a long time after thelesson begins; there is noise and disorder; and, at the start of class, more than fiveminutes are spent doing nothing. High values indicate greater problems withdisciplinary climate.

Teacher-student relations: This index was derived from students’ reports on theirlevel of agreement with the following statements: students get along well with mostteachers; most teachers are interested in students’ well-being; most of my teachersreally listen to what I have to say; if I need extra help, I will receive it from myteachers; and most of my teachers treat me fairly.

Teacher support: This index was derived from students’ reports on the frequencywith which: the teacher shows an interest in every student’s learning; the teachergives students an opportunity to express opinions; the teacher helps students withtheir work; the teacher continues teaching until the students understand; the teacherdoes a lot to help students; and, the teacher helps students with their learning.

52



School size: Number of 15-year olds in school

Schools offering extracurricular activities: This index is the proportion of studentsin a school reporting that their school offers extracurricular activities.

Negative teacher behaviour: This index was derived from principals’ reports onthe extent to which the learning by 15-year-olds was hindered by: the low expectationsof teachers; poor student-teacher relations; teachers not meeting individual students’needs; teacher absenteeism; staff resisting change; teachers being too strict withstudents; and students not being encouraged to achieve their full potential. Highvalues indicate higher levels of negative behaviour.

Student behaviour: This index summarises principals’ perceptions of the school’sdisciplinary climate by reporting the extent to which learning by 15-year-olds intheir school was hindered by: student absenteeism; disruption of classes by students;students skipping classes; students lacking respect for teachers; the use of alcoholor illegal drugs; and students intimidating or bullying other students. High valuesindicate problems with student behaviour.

Teacher morale and commitment: This index was derived from the extent towhich school principals agreed with the following statements: the morale of theteachers in this school is high; teachers work with enthusiasm; teachers take pridein this school; and teachers value academic performance.

Student-teacher ratio: This index is the ratio between the school size and the totalnumber of teachers. Part-time teachers contributed 0.5 and full-time teachers 1.0 tothe total number of teachers.

Teacher shortage: This index was derived from the principals’ views on howmuch learning by 15-year-old students was hindered by the shortage or inadequacyof teachers in the <language of assessment>, mathematics or science. High valuesindicate problems with teacher shortage.

Inadequacy of instructional resources: This index was derived based on theschool principals’ reports on the extent to which learning by 15-year-olds was hinderedby: not enough computers for instruction; lack of instructional materials in the library;lack of multi-media resources for instruction; inadequate science laboratoryequipment; and inadequate facilities for the fine arts. High values indicate a lowquality of educational resources.

Inadequacy of material resources: This index was derived from principals’ reportson the extent to which learning by 15-year-olds in their school was hindered by:poor condition of buildings; poor heating and cooling and/or lighting systems; andlack of instructional space (e.g., classrooms). High values indicate a low quality ofphysical infrastructure.

Computer availability for students: This index is the ratio of the principals’ reporton the number of computers in the school available to 15-year-olds and the numberof 15-year-olds in the school represented by the sample.

Teacher specialisation: This index was computed from 3 variables describing thepercentage of teachers in reading, mathematics, and science, respectively, teachingin their area of specialisation.

53



School autonomy: This index was derived from principals’ reports on whether ornot various aspects of school management (hiring and firing teachers, determininginitial and incremental salaries, formulating school budget, determining studentdisciplinary and assessment policy, approving student admittance, choosingtextbooks and course content, and deciding which course are offered) were a schoolresponsibility.

Teacher participation in decision making: This index was derived from principals’reports on whether or not teachers have the main responsibility, within their school,for the aspects of school management described above for school autonomy.

(Community) Aggregate parental occupational status: This variable was derivedas the average of the highest parental occupational status (ISEI) reported by allstudents in a school. The variable was then used as a characteristic for each studentin the school as a proxy for the occupational status of adults in the community.

III. What is a rural school?There are a variety of ways to define rural using the Census geographic informationon the location of schools. There is no single recommended definition. Rather, thechoice of how to define rural depends on the nature of the analysis carried out. Forthis analysis, a variety of Statistics Canada standard geographic variables wereexplored in determining what should be considered a rural school for the purposesof this analysis in addition to information provided in the PISA school questionnaires.In this report, rural schools are defined as those schools located outside large urbancentres (CMA or CA). That is, in terms of Statistics Canada definitions, rural schoolsare those in Rural and Small Towns (RST).

Census rural area: The most commonly used definition, Census urbanarea, refers to communities of at least 1000 people, and a population densityof at least 400 per square kilometre. Census rural refers to communities thatdo not meet this criterion.

Census Metropolitan Area (CMA) and Census Agglomeration (CA):These are urban cores, together with adjacent rural and urban areas thathave a high degree of economic and social integration with that urban area.A CMA is the area around an urban core with a population of at least 100,000.A CA surrounds an urban core of at least 10,000. In CMAs and CAs, landis designated as being part of an urban core, urban fringe or rural fringe.Certain rules with respect to population and density are used to make theurban designations. Outside CMAs and CAs, land is also designated asurban area (census urban) using the same rules. Area not designated urbanis considered rural (census rural).

Rural and small town (RST): This definition refers to populations livingoutside the commuting zone of larger urban centres (outside of CMAs andCAs.)

Rural postal codes: Areas serviced by rural route delivery from a postoffice or postal station. These areas are identified by the use of a 0 in thesecond position of the postal code.

54



In addition to these Statistics Canada, geographic concepts, information oncommunity size was collected in the PISA school questionnaire. Principalswere asked about the size of the community in which the school was locatedbased on population size: village or rural area, small town, town, city (of100,000 to about 1 million), close to a city of over 1 million, in a city of over1 million. It should be noted that the population size of a community doesnot necessarily provide information about the rural or urban nature of thatcommunity in terms of access to and integration with an urban centre. Acomparison of this variable with the CMA/CA definition used in this analysisshows that 32% of students in schools identified by principals as being invillages or small towns, were in fact in schools located within CensusMetropolitan Areas or Census Agglomerations.

Number of 15-year olds by location of school using definition based on Ruraland Small town (non-CMA/CA) compared to community size collected inPISA questionnaire

CMA/CA RSTPISA principal questionnaire (urban) (rural) Total

Village (less than 3,000) 6,522 26,673 33,195Small town (3,000 to 15,000) 27,711 47,552 75,263Town (15,000 to 100,000) 90,971 674 91,645City (100,000 to 1,000,000) 100,587 0 100,587City (more than 1,000,000) city centre 23,335 0 23,335City (more than 1,000,000) elsewhere 17,324 0 17,324

Total 266,450 74,899 341,349

55



APPENDIX C

Analytical methodologyThe analysis of rural-urban differences in student performance was carried out inthree phases. The first stage was an examination of a variety of individual, family,school and community characteristics to identify any significant and systematicdifferences between students in rural and urban schools. The second phase thenused a hierarchical (multilevel) regression model to identify the characteristics thatbest explain the difference between the rural and urban PISA results. Finally, avariety of school characteristics were explored to identify factors which could befurther analysed as potential tools for improving rural student performance.

Phase 1

Examination of differences in the school populations ofrural and urban schoolsIn order to answer the question why there are differences between the readingperformance of the two populations, the first phase of the analysis was to identifyother ways in which the populations differed.

The variables in this analysis came from three sources: 1) student reportedvariables on individual behaviour and family background from the PISA and YITSstudent questionnaires, 2) principal reported variables from the PISA schoolquestionnaire, and 3) community level variables from the 1996 Census aggregatedfor Census Sub-division (CSD) geographic units linked to schools using postalcode information.

Given the complex sample of the PISA assessment, group mean characteristicsat the student level were estimated using replication methods. The statistics wereestimated using 80 different weightings of the sample. The set of replication weights,produced by Westat, was designed to be consistent with PISA sampling (see below).For each statistic, the variance between the 80 different estimates was proportionalto the variance of estimation. For further information on the analytical treatment ofPISA sampling, see the PISA 2000 Technical Report (Adams and Wu, 2002). Thesemethods were implemented using the software WesVar 4.0 (2001).

At the school level, the comparison of rural and urban schools was complicatedby the nature of the school sample. Since schools were the first-stage sampling unit(see below), statistics estimated using the school design weights produced estimatesthat generalised to the population of schools that enrolled 15-year-olds, but did notgeneralise to the school environment of 15-year-olds. To illustrate the distinction,consider that a small minority of 15-year-olds repeated a grade and were attendinglower secondary schools. Similarly, a small minority of 15-year-olds were attendingschools with extremely small populations. These schools would be considered equalin estimating average school characteristics to larger, upper secondary schools thatare more typical for 15-year-olds. Using the school design weights to estimate average

56



school characteristics would produce average school characteristics that wereactually unrepresentative of the school environment experienced by the average15-year-old.

In order to remedy this inconsistency, school weights were constructed usingschool aggregations of student weights. Thus, statistics produced using these weightsare generalizable to the environments experienced by 15-year-olds. Unfortunately,this method does not come with a sample-appropriate statistical method for estimatingthe precision of statistics. Thus, in order to estimate variances to be used in statisticaltests, the aggregated weights were normalised across the sample of schools (dividedby a constant, such that the sum of the normalised weights was equal to the numberof schools sampled). Using normalised weights, variance estimates were calculatedunder the assumption of simple random sampling of schools. However, given thatthis method did not take into account any sample design complexities, it is likelythat the variance estimates were an underestimation of true variances13. Therefore,the test level of significance was changed from the typical 5% error rate to a 1%error rate. One expects that increasing the stringency of the statistical test shouldoffset the possible underestimation of variance.

Having identified many variables that distinguished rural and studentpopulations, several hypotheses were developed about possible causes for the groupdifferences across a wide variety of characteristics, particularly in our variable ofinterest, reading performance. For example, student career expectations, parentaloccupational status and education, and community-level factors all showed consistentdifferences in favour of urban communities. However, other variables, such as schoolcharacteristics and student attitudes (within-community variables) were inconsistentin terms of group differences. The systematic difference in community characteristicsbetween rural and urban schools led to an analysis of community level variables inorder to explain the systematic differences at the individual level.

Phase 2

Analysis of impact of individual and community levelvariables on rural student performance (Model Set 1)In order to test the effects of community level variables on group-averaged individualperformance, it was necessary to produce a complex model that could account forsome of the complexity of the system, while still allowing us the statistical power toestimate relationships between the variables.

The goal in this phase of the analysis was to predict the average outcomevalue for one group (focus group – rural school populations) if they had the averagecharacteristics of another group (reference group – urban school populations). Inorder to accomplish this, all explanatory variables used in the model were centredon the average value of the reference group (so the average of the transformedvalues is equal to the mean of the reference group). This results in a much moreinterpretable model, since the intercept term of the regression equation now refersto the predicted average of the population if the average of the predictor(s) for thepopulation was the reference group average. Another interpretation is that the intercept

57



term is the predicted value for all cases that have the predictor value equal to theaverage of the reference group. In the case of this analysis, the intercept would bethe mean for the urban school population.

However, it was important in this analysis to examine rural-urban differenceswithin provincial jurisdictions since education systems are governed provinciallyand it is more meaningful to compare rural school populations to urban schoolpopulations within provinces. This means that the model needed to account forprovincial differences without explaining provincial differences. In order toaccomplish this, all explanatory variables were transformed so that they did notreflect differences between provinces. This required that, for each province, allexplanatory variables were adjusted within-province. Since the objective of thisanalysis was to predict outcomes for students with average individual and communitycharacteristics of the reference groups, variables were adjusted around the mean ofthe reference group for each province. As a result, each predictor variable used inthis set of models was of the form:

kujkjk xxx **

•−= ( 1 )

where

x*jk is the final, transformed value of variable x for case j in province k that was used

in the regression models

xjk is the raw, untransformed value of the variable for case j in province k

x.uk represents the average value of the variable for the urban cases in province k

Because the predictors had been centred around the urban mean for each province,but no similar adjustment had been made to the outcome variable (readingperformance), it was necessary to include dichotomous variables that indicatedprovince, as well as an indicator variable indicating if a school was rural withineach province.

58



Figure 6

An illustration of provincial differences in predictor and outcome,with a constant relationship

Predictor

Ou

tcom

e

A

B

C

The justification for these adjustments is that many of the communitycharacteristics used to describe differences in rural and urban performance varybetween provinces. Unfortunately, the educational policies that determine outcomes,perhaps to a greater degree than community conditions, are also systematicallydifferent between provinces. It is possible that a province with systematically highersocio-economic conditions will have systematically different educational policies.This confound is illustrated in Figure 1, which presents 3 hypothetical provinces,each with systematically different locations in terms of both predictor and outcome.The traditional scatterplots are represented here as ovals. The relationship betweenpredictor and outcome is the same in each province, indicated by the parallel linesbisecting each oval. However, since the provinces are systematically different interms of these variables, the observed effect if all the provinces were consideredsimultaneously would be much different (shown here as the solid line). However,this overestimated relationship would be an artefact of systematic differences betweenprovinces, instead of the actual relationship between variables.

By centring the predictor within-province, the effect of provincial membershipis negated (see Figure 2). Since the average value of the predictor is now identicalfor each province, the distributions have been aligned horizontally. The distributionsremain displaced vertically, which has the effect of attenuating the observedrelationship, shown here by the flatter solid line. In order to account for thisdisplacement, it is necessary to include variables that account for provincialmembership. Since the distributions are already aligned according to the predictor,the variable accounting for provincial membership only describes the differencesbetween provincial means. The resulting distributions, with both within-provincecentring and provincial indicators, are shown in Figure 3. The distributions havebeen effectively overlain on top of each other, and observed relationship is nowequal to the actual relationship within each province.

59



Figure 7

The effect of centring variables within-province

Predictor

Out

com

e

A

B

C

Figure 8

The combined effect of centring and use of provincial indicators

Out

com

e

Predictor

A, B and C

60



For every model, there were a total of 9 province variables (10-1) and 10rural-urban variables (1 for each province). The provincial indicators were dummy-coded (for an explanation of dummy coding, see Cohen and Cohen, 1983,pp. 183-220) against the reference group of Alberta, and the rural-urban indicatorswere dummy coded against the urban group within each province. This meant thatthe intercept term for the regression equation represented the average performanceof urban students in Alberta. The regression coefficient for each indicator variablerepresented an adjustment from the urban Alberta average. The averages of studentsin urban school in other provinces would be adjusted by the value of the provincialindicator coefficient, while averages of students in rural school would be adjustedby both the provincial indicator coefficient and the rural-within-province indicatorcoefficient. These adjustments to the model allowed us to account for the differencesbetween provinces without explaining them. For example, since the differencebetween average Alberta urban performance and average Newfoundland andLabrador urban performance is perfectly accounted for by the variable indicatingwhich schools are in Newfoundland and Labrador, we can account for theinterprovincial variation in performance without explicitly defining why it exists.

If we represent the vector of average performance within each school as B0,the vector of provincial indicators as P and the vector of within-province ruralindicators as R, the basic model that accounts for the differences between rural andurban groups in the different provinces is:

USGRGPGGB ll ++++= ∑ 100201000 ( 2 )

where G00

is the average performance of students in urban Alberta schools, G01

isthe vector of differences between the urban performance in Alberta and the otherprovinces, G

02 is the vectors of values describing the difference between rural and

urban performance in each province, and U is a vector of school residuals fromtheir predicted provincial-geographic group means. The summation term across G

0lrepresents the combined effects of any community characteristics, S1l, used in themodel. However, in order to account for the wide variation within each school interms of performance, the model above (2) was combined with a student level model,Equation (3), that describes individual students’ performance, A, as a sum of theschool average performance, B0, and within-school residuals for each student, E.The resulting model obtained by substituting (2) into (3), accounts for both differencesbetween school averages and between individual students within each school (4).

EBA += 0 ( 3 )

ERSGRGPGGA ll +++++= ∑ 10020100 ( 4 )

Another consideration in the analysis of macro level variables on groups ofindividuals is that there is a risk of confounding group characteristics with theaggregate effect of individual characteristics. For example, a group characteristicmay have a relationship to the average group outcome, but this relationship may

61



simply be the aggregate effect of correlated characteristics of individuals within thegroup. In other words, if average income seems to be related to average performance,it is important to be certain that this is not simply because individual income isrelated to individual performance. In order to control for this situation, all modelsused in this analysis controlled for the socio-economic conditions of individuals.Again, these variables were centred within-province around the average of thereference group. This required an adjustment to Equation (2), so that the jth elementof B

0 represents the predicted average performance of students in school j if all the

students in the school had families with average urban characteristics. The newindividual-level model is represented as:

EXBBA pp ++= ∑0 ( 5 )

where the Xp represents a matrix of individual variables describing individual familysocio-economic characteristics, centred on the urban average, and B

p is a matrix of

the regression coefficients of these variables onto individual performance, estimatedwithin each school. The two individual level variables used to describe family socio-economic characteristics were the highest parental occupational status and the highestparental educational attainment. The final model used to describe differences instudent performance, combining (2) and (5) is:

ERXBSGRGPGGA ppll ++++++= ∑∑ 10020100 ( 6 )

The elements in G00

, G01

, and G02

now describe the predicted averageperformance of the relevant groups if the average characteristics (S1l and Xp) ofeach group were equal to the within-province average urban characteristics. If allS1

l and X

p are empty, which is equivalent to modelling the group differences without

any predictor variables, combining the elements in G00

, G01

, and G02

will producethe observed averages for each group. Since the urban schools in Alberta had thehighest average performance of any other of the provincial geographic groups, theelements in G

01 will all be negative, describing the deviation of each province’s

urban average from Alberta’s. Furthermore, since the urban averages were higherthan rural in all provinces, the elements in G02 will also be negative, describing thedeviation of each provincial rural average from the provincial urban average. Byincreasing the number of predictors, we are attempting to reduce the absolute valuesin G

02 by explaining the differences between rural and urban performance. Since

the reference urban groups already have average urban characteristics, elements inG00 and G01 will not change in value. However, if the characteristic, S1l, has somepower in explaining the difference in performance between students in rural versusurban schools, then the values in G

02 will become less negative. In general, as the

elements of G02 increase in relative value, we are explaining more of the differencebetween rural and urban performance. If the elements become positive, then itsuggests that students in rural schools are performing better than predicted, giventhe conditions we have modelled. The predicted performance of rural students inall provinces is calculated by adding the corresponding elements in G01P and G02Rto G00.

62



Phase 3

Analysis of the potential impact of school-levelvariables on rural student performance (Model Set 2)This last phase of the analysis took into consideration that the community contextof schools is not easily changed. However, given that schools do exist within specificcontext(s), it is useful to identify school variables which were not reported at thehighest levels and which were associated with high student achievement. Thismodel set looked specifically at the predicted performance of rural students, giventhe particular characteristics of rural communities and their effects noted in ModelSet 1.

When talking about the effects of school conditions, there is a risk that thereare a variety of confounding variables. In particular, because the population forPISA is 15-year-olds and is not school grade specific, many schools in the PISAsample were early-secondary or middle schools where the 15-year-olds were likelystudents who had been held back a grade at some point in their academic history.Thus, the average of these schools is expected to be lower, simply because of thesystematic differences in their sample of students. In order to control for this spuriousrelationship, a variable describing each student’s school grade was introduced. Themodel appeared as follows:

ERXBGRADEBSGRGPGGA ppmm +++++++= ∑∑ 10020100 2 ( 7 )

where GRADE describes the grade of a student, a discrete whole number variable,centred on 10. Thus, the elements in G

00, G

01, and G

0214 now describe the predicted

performance of grade 10 students in each provincial geographic region. S2m and Bprepresent vectors of a school socio-economic characteristics and individual socio-economic characteristics, respectively, centred on the rural average for each province.This model set includes all predictors identified in Model Set 1. The G

00 coefficient

describes the predicted average performance of rural grade 10 students in Alberta.Since G

02 indicates rural or urban location of each school, the summation term

across all S2m and B

p vectors does not affect the predicted provincial rural averages.

These variables are included in order to reduce the chances of observing spuriousrelationships between performance and the variables tested in this second modelset.

Using this base model, several school policy variables were introduced inorder to determine their effect on student performance, while controlling for theimportant community variables identified in Model Set 1. In order to identifyimportant variables, the variables tested in the model were centred on their maximumvalues (in other words, the variables described the difference between the actualvalue and the desirable value):

maximumjj xxx −=*( 8 )

63



where

x* is the final, transformed value of variable x for case j that was used in the regressionmodels

x is the raw, untransformed value of the variable for case j

xmaximum

represents the maximum value of the variable

For certain variables, such as the number of 15-year-olds in a school, the relationshipappeared to be non-linear. For variables with non-linear relationships, the followingtransformation was performed:

( )22*maximumjj xxx −= ( 9 )

This 2nd order term was combined with the 1st order term in order to estimate thenon-linear effect of the variables.

The following table presents the values used as maximum for each variable.These values were determined by examining the bivariate relationships andscatterplots of each variable with school average performance. For linearrelationships, the maximum value was fixed as the minimum or maximum observedvalue for a variable, depending on whether the bivariate relationship between thevariable and performance was negative or positive, respectively. For non-linearrelationships, maximum values were specified based on literature, if available, andobserved maxima in scatterplots.

Variable Maximum Value

Teacher support Maximum observed score 1.61

Disciplinary climate Minimum observed score -1.54

School activities Maximum proportion of studentsreporting that school offers activities 1.00

Achievement press Maximum observed score 1.38

Professional development Maximum % of teachers withrecent professional development 100%

Instructional hours Minimum observed total hours of instruction 100

Student-teacher ratio 25 students per teacher 25

Teacher behaviour problems Minimum observed score -2.41

Student behaviour problems Minimum observed score -2.61

Teacher morale Maximum observed score 1.78

School autonomy Maximum observed score 1.72

Teacher participation Maximum observed score 3.70

Teacher specialisation Maximum proportion of teachersteaching in their specialisation 1.00

Number of 15-year-olds Various values between 100 and 40015-year-olds were tested 100-400

Standardised testing Using standardised tests either less thanor more than 2 times per year na

64



Thus, the elements of G00, G01, and G02 in the full model sum to produce thepredicted average performance for rural students in each province if all schoolswere to have maximum values on the predictors included, while controlling forsocio-economic characteristics. For example, if the Alberta rural average forproportion of teachers specialising in their instructional content area were 80%, themodel predicts what average rural performance would be if that proportion were100%. This final model is defined as

ERXBGRADEBSGSGRGPGGA ppnnmm ++++++++= ∑∑∑ 100020100 32

( 10 )

where S3n represents the difference between a school variable and its maximum

value. This model was used to identify variables which were not reported at thehighest levels and, b) predict a significant change in student performance if theywere given maximum values.

Constraints on this analysisThe first consideration in modelling these data was that the issue of estimatingstandard deviations from a complex sample had not yet been resolved. As a result,the traditional interpretation of effect sizes, which rely on accurate estimation ofpopulation standard deviations, to interpret the relative relationships betweenvariables, could not be done. Thus, the analysis was constrained to fitting modelsand predicting values according to the fitted models, rather than reporting andinterpreting effect sizes. An unexpected benefit of this method is that the results aremuch more communicable to a lay audience, since predicted averages are closer tocommon experience than are standard deviations and regression coefficients.

All coefficients in Phases 2 and 3 were estimated using hierarchical ordinaryleast squares (OLS) estimation for both within-school and between-school effects.The models were replicated across the 5 plausible values describing the posteriordensity function for each individual’s reading performance, and the results of thefive analyses were averaged to produce the final reported estimates. Within school(individual) effects were allowed to vary randomly between schools. The softwareused for estimation of coefficients was HLM 5 (Raudenbush, S., Bryk, T., &Congdon, R, 2000). Although this software also produces optimal estimates usingBayesian estimation for individual level effects, vector Bi, a consequence of thistechnique is that school averages shrink towards the grand average (see Chapter 3,Bryk and Raudenbush, 1992). As a result, the predicted means for provinces wouldno longer be equivalent to the observed means, which have already been published.The potential cost of using the OLS estimates is instability in the estimation ofpredicted school averages, given the individual level model. Analysis of thedistribution of OLS predicted averages indicates a greater variability in predictedmeans than was observed in the actual data. This variability disappears when within-school effects are fixed, rather than left to vary. However, the predicted provincial/geographic averages (the only reported statistics from these sets of models) arestable whether the within school effects are treated as fixed or not. Thus, the trade-off between consistency and optimisation was decided in favour of consistency,and the OLS estimates were used rather than the Bayesian estimates. If individual-

65



level effects varied significantly between schools according to a chi-squared test ofsignificance, the effects were left to vary; otherwise, they were treated as fixed. Allstatistical tests in these models were at the 0.05 level.

Other algorithms for modelling these effects were considered, in particular,the disaggregation of school and community characteristics onto individual students.Aside from the violation of the assumption of independence, given the clustering ofstudents in schools, the results produced using this method were relatively unstable,in that the urban averages, which should remain constant across models given thedata transformations described above, varied noticeably.

Both model sets rely on several assumptions, particularly that the variablesbeing measured (or their second-order transformations) have a linear relationshipwith performance. This is a basic assumption of regression, and the tenability of itwas examined through scatterplots. It was also assumed that the slope of theserelationships did not differ between provinces (see explanation of Model Set 1,above). This assumption provides the analysis with greater power to estimate schooleffects. Although there is no reason to believe that effects should differ betweenprovinces, excepting random sample-dependent fluctuations, this assumption wasuntested in this analysis.

A major limitation of this analysis is the limited number and descriptive powerof the variables available. It is possible that other variables describing socio-economicconditions or social capital would produce a clearer picture of community effects.Likewise, many aspects of school environments, such a principal leadership, schoolclimate, and community-school interaction, were unavailable for this analysis. Asmore data are released as part of the ongoing YITS/PISA projects, better indicatorsof school and community variables will become available, enriching this type ofanalysis.

References

��

�� ! ��"��#�� !��$$��%$�$��$�!��

��&��"��'��%��())��" �� # ��$�� %��! �� &�� $��*"��&��%�+"

��'"�� '"�� (),-�� $�� '�� ()�� %��! �� *��*"��.��&��

��"��'��%��&��/��+��01 �2�0�"��!'�!��1�"��*��"�� "��+�3��4�$"��5$��"6�1�!��%!�"$�5�!�%�5$��"�7$"��$��7!

�"�8��9��3��4�$"��5$��"6��(��'�+$��"�$�$

66



APPENDIX D

Survey Concepts, Methodology and Data QualityThe following information should be used to ensure a clear understanding of thebasic concepts that define the data provided in this report, of the underlyingmethodology of the survey, and of key aspects of the data quality. This informationwill provide a better understanding of the strengths and limitations of the data, andof how they can be effectively used and analysed. The information may be ofparticular importance when making comparisons with data from other surveys orsources of information, and in drawing conclusions regarding change over time,differences between geographic areas and differences among sub-groups of thetarget population.

Survey objectives: The Youth in Transition Survey (YITS) is a new Canadianlongitudinal survey designed to examine the major transitions in young people’slives, particularly with respect to education, training and work. Information on theskills of the 15-year-olds who participated in YITS was also collected as part of theProgramme for International Assessment (PISA).

Target populationThe target population of the 15-year-old cohort of YITS/PISA are all youth whowere 15-years-old as of December 31, 1999 who were enrolled in an educationalinstitution in Canada.

Sample designThe sample for PISA/YITS was selected in two-stages. In the first stage, the framewas a list of all provincially governed public and private schools where studentsborn in 1983 were enrolled for the 1998/1999 school year. Strata of schools werecreated to ensure adequate coverage of all ten provinces, as well as of minoritylanguage school systems in Nova Scotia, New Brunswick, Quebec, Ontario, andManitoba. The size of the school, measured by the enrolment of students born in1983, was the final stratification variable. In strata containing the largest schools, allschools were selected, whereas in other strata representing schools of size 35 ormore, schools were randomly sampled in proportion to the enrolment of studentsborn in 1983.

Within strata of schools smaller than size 35, schools were randomly sampledwith equal probability. In the second stage, students were sampled randomly from alist of the 15-year-olds enrolled in the school. In most strata a maximum of 35students were sampled, but school sample sizes greater than 35 were required insome province/language classes to meet data quality requirements. In schools withfewer than 35 eligible students, all were selected.

Stratification: In Canada, the selection of schools was carried out to ensureadequate coverage of all ten provinces, as well as of minority language schoolsystems in Nova Scotia, New Brunswick, Quebec, Ontario, and Manitoba.

67



Within the context of the sampling standards of PISA, some schools andsome students were excluded from the study. Schools in the Yukon, NorthwestTerritories and Nunavut (0.43% of the target population), as well as schools onIndian reserves (0.73%) were not included in the sampling frame. In addition studentswho were mentally or physically disabled in such a way that they could not performin the PISA assessment and non-native speakers with less than one year of instructionin the language of assessment, as well as schools that teach only these students,could be excluded from the study. Schools with enrolment of fewer than 3 studentsin the target population were also excluded from the study.

Total population 15-years old 403,803

Total enrolled population 396,423

Total in target population 391,788

School-level exclusions 2,035

Percentage of school-level exclusions 0.52%

Number of participating students 29,687

Weighted number of participating students 348,481

Number of excluded students 1,584

Weighted number of excluded students 16,197

Student (within-school) exclusion rate 4.44%

Overall exclusion rate 4.94%

��

Collection

PISA AssessmentThe PISA 2000 survey included a direct assessment of students’ skills throughreading, mathematics and science tests. A total of about seven hours of test itemswere administered, with each student taking a two-hour-long assessment consistingof different combinations of test items. The assessment focused mainly on reading,with the reading test giving three sub-test scores labelled retrieving information,interpreting and reflecting. Mathematics and science each had only a single score.In addition, as minor domains, fewer mathematics and science items were includedand these items were administered to a random sub-sample of PISA participantswithin each participating school.

In Canada, students were assessed in English or French according to theirmain language of instruction as determined by the school.

68



Student, school and parent questionnairesStudents also completed a 20-minute questionnaire focussing on factors contributingto student performance and a three-minute questionnaire focussing on informationtechnology. In addition, PISA 2000 included a questionnaire, which was administeredto school principals, to collect information about the characteristics of participatingschools.

A 30-minute self-completed contextual questionnaire from the Youth inTransition Survey was administered simultaneously to students in order to collectmore information on their school experiences, their work activities and theirrelationships with others. A 30-minute interview was also conducted with a parentof each student.

For further information on the technical background of PISA (includingresponse rates), see Annex A of the international OECD report Knowledge andSkills for Life - First results from the OECD Programme for International StudentAssessment, and the PISA 2000 Technical Report. These reports are available onthe Internet at www.pisa.oecd.org.

69



Endnotes

1 According to the OECD preliminary results of the Programme for International StudentAssessment, Canada is one of the countries where the range in student achievement is dueless to differences between schools, than to differences in students within schools. Thecountries, like Canada, where these “between-school” differences are relatively small,tend to be the highest performing countries. The fact that Canadian schools are relativelyuniform is therefore seen as one of the reasons why Canada performed so well compared toother countries. (Knowledge and Skills for Life, OECD, Paris, 2001)

2 The location of schools rather than students’ homes was used for this analysis because oneimportant aspect of this study is to identify whether differences in the schools themselvesare important factors in urban – rural differences. While it is also important to understandthe location of students’ homes, this information was not available for this analysis.

3 The PISA 2000 assessment tested students in reading, mathematics and science literacy.However, only a sub-sample of students was tested in mathematics and science. As a result,a full analysis of mathematics and science literacy in rural and urban student populationswas not possible due to restrictions of sample size. For example, partly as a result ofsample size, differences in the mathematics and science performance of rural and urbanstudents were not statistically significant in most provinces. In mathematics, only inNewfoundland and Labrador was there a significant difference between the performanceof rural and urban students. In science, rural-urban gaps were significant only inNewfoundland and Labrador, New Brunswick and Alberta.

4 In Newfoundland and Labrador, Prince Edward Island, New Brunswick and Alberta, thedifference in reading performance of rural and urban students was statistically significantwith a 99% level of confidence. In Ontario, the rural-urban gap was significant with a 95%level of confidence.

5 The PISA assessment was implemented in each province through provincial ministries ofeducation. On-reserve schools that do not fall within provincial jurisdiction were notincluded in the PISA study. As a result, there will be some under-coverage of rural studentpopulations.

6 Family socio-economic background was derived from student responses regarding parentaloccupation using the International Socio-economic Index of Occupational Status (seeAppendix B for definition). This scale was also used to derive occupational status forstudent career aspirations.

7 In Alberta and Ontario, the rural-urban difference in cultural activities is significant witha 95% level of confidence, but not at the 99% level generally used to measure statisticalsignificance in this paper.

8 In spite of these variations in the amount of time spent getting to school, there was norelationship between transportation times and reading performance. In fact, additionalanalysis of this data also showed that there was no relationship between transportationtimes and participation in extracurricular activities or a student’s sense of belonging.However, issues surrounding the transportation of students to and from school are complexand extend beyond the impact on academic performance. Concerns over studenttransportation also involve matters such as student safety as well as the costs oftransportation relative to school budgets. This analysis is unable to fully explore all of theelements of this issue, nor does it consider the possible impact on students of transportationtimes in earlier grades.

70



9 It should be noted that this does not mean that many of these youth will not end up goingto university, especially given the opportunities for university preparation and transferprograms in colleges and CEGEPs. As the 18 to 20-year old cohort of YITS are still at thebeginning of their studies, it is still too early to determine the highest level of educationthey will attain.

10 For the most part, the schools in the PISA sample are high schools or include high schoolgrades. However, because the study is representative of 15-year-olds, it includes a mix ofschools and grades as some 15-year-olds can be found in earlier grades in middle or juniorhigh schools (as well as in more advanced grades in high schools). For this reason, schoolcharacteristics cannot be interpreted as characteristics of high schools, or even of schoolsincluding high school grades.

11 In British Columbia, the average income of individuals was higher in the urbancommunities, but this was not a statistically significant difference.

12 This analysis was focused on Grade 10 students in rural schools.13 These considerations were weighed against the fact that the sampling fraction of the

number of schools in many provinces approached a census. Although sampling fractionwas not considered here for consistency with previous PISA analyses, adjustment for thesampling fraction of schools would have reduced the variance estimates for manyjurisdictions to be near zero.

14 Although it did not change the properties of the model, the elements of G02 were reversed,such that the intercept G00 refers to the average predicted grade 10 performance in ruralAlberta. This adjustment simplified the interpretation of the final model.

71



Culture, Tourism and theCentre for Education StatisticsResearch PapersCumulative Index

Statistics Canada’s Division of Culture, Tourism and the Centre for EducationStatistics develops surveys, provides statistics and conducts research and analysisrelevant to current issues in its three areas of responsibility.

The Culture Statistics Program creates and disseminates timely and comprehensiveinformation on the culture sector in Canada. The program manages a dozen regularcensus surveys and databanks to produce data that support policy decision andprogram management requirements. Issues include the economic impact of culture,the consumption of culture goods and services, government, personal and corporatespending on culture, the culture labour market, and international trade of culturegoods and services. Its analytical output appears in the flagship publication Focuson Culture (www.statcan.ca/english/IPS/Data/87-004-XIE.htm) and in Arts, cultureand recreation – Research papers.

The Tourism Statistics Program provides information on domestic and internationaltourism. The program covers the Canadian Travel Survey and the InternationalTravel Survey. Together, these surveys shed light on the volume and characteristicsof trips and travellers to, from and within Canada. Its analytical output appears inthe flagship publication Travel-log (www.statcan.ca/english/IPS/Data/87-003-XIE.htm) and in Travel and tourism – Research papers.

The Centre for Education Statistics develops and delivers a comprehensiveprogram of pan-Canadian education statistics and analysis in order to support policydecisions and program management, and to ensure that accurate and relevantinformation concerning education is available to the Canadian public and to othereducational stakeholders. The Centre conducts fifteen institutional and over tenhousehold education surveys. Its analytical output appears in the flagship publicationEducation quarterly review (www.statcan.ca/english/IPS/Data/81-003-XIE.htm),in various monographs and in Education, skills and learning – Research papers(www.statcan.ca/english/IPS/Data/81-595-MIE.htm).




Statistics StatistiqueCanada Canada

Human Resources Développement des ressourcesDevelopment Canada humaines Canada

Council of Ministers of Education, CanadaConseil des ministres de l’Éducation (Canada)

72



Following is a cumulative index of Culture, Tourism and Education researchpapers published to date

Arts, culture and recreation – Research papers

Forthcoming

Travel and tourism – Research papers

Forthcoming

Education, skills and learning – Research papers

81-595-MIE2002001 Understanding the rural-urban reading gap

understanding the rural - urban reading gap · understanding the rural-urban reading gap catalogue...

Documents